Aerial image

This communication presents an approach to generate high resolution orthophotos and a digital surface model combining RPAS imagery and ground control support derived from aerial and publicly accessible data instead of organizing a specific surveying campaign. The scope of the presented research is the multi-resolution geospatial data management and this research seeks to contribute to this topic by assuring geo-referencing consistency between RPAS imagery and aerial and historical orthophotos. The approach has been experimentally tested and validated in the Roman city of Pollentia (Alcúdia, Mallorca, Spain) and might be further applied to any other archeological site where aerial imagery and auxiliary data are available.

Realistic images of terrain can be created by texture mapping aerial photographs onto elevation data. However, lighting and shadows in the original photographs are a problem when multiple aerial photos are merged, or when the sun direction used for the generated image is different from the original aerial photograph. Our goal is to generate a terrain surface reflectance map from an aerial photograph. This will be used to simulate a new aerial photograph with different lighting and viewing directions. We are concentrating on forest and mountain areas for use in ecosystem management and modeling applications. One key to generating realistic new images is finding shadows cast by the terrain in the original aerial photographs. Digital elevation values are uncertain due to vertical imprecision in the measurement process and coarse sample spacing. This uncertainty can result in substantial differences between calculated shadow positions and actual shadow locations in photographs when conventional shadow location methods are used. We have developed a technique which takes into account this uncertainty to determine the probability that any given point on the terrain is in sunlight or shadow in the aerial photograph. This shadow probability is crucial to accurate calculation of the reflectance of terrain points corresponding to the pixels of the original aerial photograph.

Realistic images of terrain can be created by texture mapping aerial photographs onto elevation data. However, lighting and shadows in the original photographs are a problem when multiple aerial photos are merged, or when the sun direction used for the generated image is different from the original aerial photograph. Our goal is to generate a terrain surface reflectance map from an aerial photograph. This will be used to simulate a new aerial photograph with different lighting and viewing directions. We are concentrating on forest and mountain areas for use in ecosystem management and modeling applications. One key to generating realistic new images is finding shadows cast by the terrain in the original aerial photographs. Digital elevation values are uncertain due to vertical imprecision in the measurement process and coarse sample spacing. This uncertainty can result in substantial differences between calculated shadow positions and actual shadow locations in photographs when conventional shadow location methods are used. We have developed a technique which takes into account this uncertainty to determine the probability that any given point on the terrain is in sunlight or shadow in the aerial photograph. This shadow probability is crucial to accurate calculation of the reflectance of terrain points corresponding to the pixels of the original aerial photograph.

As the industry moves towards sub-65nm technology nodes, the mask inspection, with increased sensitivity and shrinking critical defect size, catches more and more nuisance and false defects. Increased defect counts pose great challenges in the post inspection defect classification and disposition: which defect is real defect, and among the real defects, which defect should be repaired and how to verify the post-repair defects. In this paper, we address the challenges in mask defect verification and disposition, in particular, in post repair defect verification by an efficient methodology, using SEM mask defect images, and optical inspection mask defects images (only for verification of phase and transmission related defects). We will demonstrate the flow using programmed mask defects in sub-65nm technology node design. In total 20 types of defects were designed including defects found in typical real circuit environments with 30 different sizes designed for each type. The SEM image was taken for each programmed defect after the test mask was made. Selected defects were repaired and SEM images from the test mask were taken again. Wafers were printed with the test mask before and after repair as defect printability references. A software tool SMDD-Simulation based Mask Defect Disposition-has been used in this study. The software is used to extract edges from the mask SEM images and convert them into polygons to save in GDSII format. Then, the converted polygons from the SEM images were filled with the correct tone to form mask patterns and were merged back into the original GDSII design file. This merge is for the purpose of contour simulation-since normally the SEM images cover only small area (~1 μm) and accurate simulation requires including larger area of optical proximity effect. With lithography process model, the resist contour of area of interest (AOI-the area surrounding a mask defect) can be simulated. If such complicated model is not available, a simple optical model can be used to get simulated aerial image intensity in the AOI. With built-in contour analysis functions, the SMDD software can easily compare the contour (or intensity) differences between defect pattern and normal pattern. With user provided judging criteria, this software can be easily disposition the defect based on contour comparison. In addition, process sensitivity properties, like MEEF and NILS, can be readily obtained in the AOI with a lithography model, which will make mask defect disposition criteria more intelligent.

The current study sets out to develop an empirical line method (ELM) radiometric calibration framework for the reduction of atmospheric contributions in unmanned aerial systems (UAS) imagery and for the production of scaled remote sensing reflectance imagery. Using a MicaSense RedEdge camera flown on a custom-built octocopter, the research reported herein finds that atmospheric contributions have an important impact on UAS imagery. Data collected over the Lower Pearl River Estuary in Mississippi during five week-long missions covering a wide range of environmental conditions were used to develop and test an ELM radiometric calibration framework designed for the reduction of atmospheric contributions from UAS imagery in studies with limited site accessibility or data acquisition time constraints. The ELM radiometric calibration framework was developed specifically for water-based operations and the efficacy of using generalized study area calibration equations averaged across variable illumination and atmospheric conditions was assessed. The framework was effective in reducing atmospheric and other external contributions in UAS imagery. Unique to the proposed radiometric calibration framework is the radiance-to-reflectance conversion conducted externally from the calibration equations which allows for the normalization of illumination independent from the time of UAS image acquisition and from the time of calibration equation development. While image-by-image calibrations are still preferred for high accuracy applications, this paper provides an ELM radiometric calibration framework that can be used as a time-effective calibration technique to reduce errors in UAS imagery in situations with limited site accessibility or data acquisition constraints.

This paper focuses on the potential of an integrated approach using aerial LiDAR, aerial and terrestrial photogrammetry, terrestrial laser scanning, and archaeological survey to detect the presence and configuration of lost medieval settlements under canopy. This approach was applied to the site of Altanum (Calabria, Italy), on the hill of Sant’Eusebio, completely covered by vegetation. Altanum was a large fortified settlement characterised by a long occupation, especially during the Byzantine and Norman-Swabian periods. The activity began by carrying out a LiDAR survey of the whole hill. The acquired LiDAR data were processed and filtered in order to obtain a DFM (Digital Feature Model) useful for the identification of features of archaeological interest. Several enhancement techniques were performed on DFM to increase the visibility of archaeological features. The features thus identified were subsequently surveyed through the use of terrestrial and aerial photogrammetry integrate...

Coastal structures induce important impacts on the beaches and surf zone morphodynamic processes. Besides, they influence the refraction/diffraction patterns, limiting the development of long-shore currents, rips and rip feeder currents (Komar, 1998 and Short, 1999). The Portuguese coastal zone has several coastal structures, build on an attempt to minimize the coastal erosion process (groins) or port development (breakwaters). In general

Absolute localization of a flying UAV on its own in a global-navigation-satellite-system (GNSS)-denied environment is always a challenge. In this paper, we present a landmark-based approach where a UAV is automatically locked into the landmark scene shown in a georeferenced image via a feedback control loop, which is driven by the output of an aerial image registration. To pursue a real-time application, we design and implement a speeded-up-robust-features (SURF)based image registration algorithm that focuses efficiency and robustness under a 2D geometric transformation. A linear UAV controller with signals of four degrees of freedom is derived from the estimated transformation matrix. The approach is validated in a virtual simulation environment, with experimental results demonstrating the effectiveness and robustness of the proposed UAV self-localization system.

Identification  of  agricultural  areas  in  remote  sensing  technology  is  needed  for  the development  of  agricultural  areas.  The  image  of  the  agricultural  area  in  this  study  uses  an Unmanned Air Vehicle (UAV). The results of images taken from a height of 100 meters on the ground will be stored and processed into one image. UAV technology that supports this research is  expected  to  help  remote  sensing  in  real  time.  For  the  current  study,  measurements  in agricultural areas are related to some fragmented images. This article creates a beautiful view of the agricultural region. The author focuses on automatic image milling methods with detection- based image matching and description of patented local features from the dataset. The features method applied is based on speeded up robust features (SURF). The method of matching images and verification results is carried out. The result will create a 2-D spatial reference that starts the panorama size. This pap...

Many supervised and unsupervised algorithms for the automated and reliable classification of large regions using high resolution data have been presented in the remote sensing community in the last decades. Most of these approaches exploit a single input data: high resolution orthophotos or 3D point clouds. An increasing number of contributions has more recently exploited the combined use of orthophotos and LiDAR DSM taking advantage from the complementarity of these inputs. Nevertheless, very few applications have considered the use of overlapping multi-view images on the same area for classification. In this paper the first tests to investigate this classification architecture are presented. Different typologies of DSM and orthophoto as well as a variable number of images on the same area have been considered in the experiments. The preliminary results on the two available test areas will be shown, in order to draw the first conclusions on this approach and discuss the further developments of this research.

Vehicle detection and classification are the most significant and challenging activities of an intelligent traffic monitoring system. Traditional methods are highly computationally expensive and also impose restrictions when the mode of data collection changes. This research proposes a new approach for vehicle detection and classification over aerial image sequences. The proposed model consists of five stages. All of the images are preprocessed in the first stage to reduce noise and raise the brightness level. The foreground items are then extracted from these images using segmentation. The segmented images are then passed onto the YOLOv8 algorithm to detect and locate vehicles in each image. The feature extraction phase is then applied to the detected vehicles. The extracted feature involves Scale Invariant Feature Transform (SIFT), Oriented FAST and Rotated BRIEF (ORB), and KAZE features. For classification, we used the Deep Belief Network (DBN) classifier. Based on classification...

in Krakow in 1978. In 1999-2004 he was General Surveyor of Malopolska Region. Currently he is Head of Department of Geoinformation, Photogrammetry and Remote Sensing of Environment in AGH. His research fields are: radiometric quality of photogrammetric images, integration of photogrammetry and LIDAR, usage of orthoimage and stereo orthoimage in GIS environment.

Using aerial inspection techniques in farmlands can yield vital data instrumental in mitigating various impediments to optimizing farming practices. Farmland anomalies (standing water and clusters of weeds) can impede farming practices, leading to the improper utilization of farmland and the disruption of agricultural development. Utilizing Unmanned Aerial Vehicles (UAVs) for remote sensing is a highly effective method for obtaining extensive imagery of farmland. Visual data analytics in the context of automatic pattern recognition from collected data is valuable for advancing Deep Learning (DL) -assisted farming models. This approach shows significant potential in enhancing agricultural productivity by effectively capturing crop patterns and identifying anomalies in farmland. Furthermore, it offers prospective solutions to address the inherent barriers farmers encounter. This study introduces a novel framework, namely the hybrid Convolutional Neural Networks and Long Short-Term Memory (HCNN-LSTM), which aims to detect anomalies in farmland using images obtained from UAVs automatically. The system employs a Convolutional Neural Network (CNN) for deep feature extraction, while Long Short-Term Memory (LSTM) is utilized for the detection task, leveraging the extracted features. By integrating these two Deep Learning (DL) architectures, the system attains an extensive knowledge of farm conditions, facilitating the timely identification of irregularities such as the presence of water, clusters of weeds, nutrient deficit, and crop disease. The proposed methodology is trained and evaluated using the Agriculture-Vision challenge database. The results obtained from the experiment demonstrate that the proposed system has achieved a high level of accuracy, with a value of 99.7%, confirming the effectiveness of the proposed approach.

Achieving accurate orthorectification is a major constraint to upscaling the use of historical aerial imagery for 20th century change detection. This paper presents a series of aerial triangulation bundle block adjustment post-processing model tests to determine the planimetric accuracy obtainable for large historical image blocks, which inherently contain film deformations e.g. warpage and shrinkage. Photo jobs with a range of photo numbers (12-237 images) and area coverage (>33 000-~150 000ha) were included. Self-calibration with 44-parameters and fixing the model to the ground control network achieved the highest final planimetric accuracy (total root mean square error [RMSE] ranging from 18.6 – 25.7px [7.1 – 21.8m] at ground control points and 20.5 – 27.7px [7.8 – 23.5m] at checkpoints). Allowing movement in the model by increasing standard deviations at ground control points or automating the removal of blunders in the model significantly reduced final planimetric accuracy. ...

Field studies were conducted in 1998 and 1999 in Livingston Field at Perthshire Farm, Bolivar County which is located in west-central Mississippi along the Mississippi River. It is a 162 ha field and had a 2-m elevation range. The dominant soil series of the field are Commerce silt loam, Robinsonville sandy loam and Souva loam. The objectives of the study were to 1) compare GOSSYM simulated yield with actual yield, 2) study spatial and temporal pattern of cotton crop across two growing seasons using multispectral imagery, 3) predict field based lint yield from remote sensed data, and determine age of the crop most suitable for aerial image acquisition in predicting yield and / or discriminating differences in cotton growth. Two transects were selected for GOSSYM study, each containing twelve sites. A 1-m length of single row plot was established at each profile. Plant mapping were done five times in 1998 and seven times in 1999 growing seasons. GOSSYM simulation runs were made for each profile and compared with actual crop parameters using root mean square error (RMSE). Results based on averaging common soil mapping units indicate that GOSSYM accuracy in predicting yield varied from 0.45 bales acre -1 to 0.61 bales acre -1 . To monitor and estimate the biophysical condition of the cotton crop, airborne multispectral images were acquired on 10 dates in 1998 and 17 dates in 1999 from April to September. In both years site-specific yield and normalized difference vegetation index (NDVI) were significantly (p < 0.0001) correlated in the month of July. NDVI curves of different sites in 1999 showed least distinctiveness due to somewhat wetter weather conditions as compared to drier weather in 1998. Crop growing in or near the drainage areas were low in NDVI and lint yield. Multispectral images acquired between ~ 300 -600 growing degree days (GDD60) may be useful decision tools for replanting certain areas of the field, especially in dry weather conditions when variability in crop growth pattern is enhanced due to spatial variability in soil texture, which dictates soil moisture holding capacity and its release to plants. Results suggest that 2-3 multispectral images acquired between 800 and 1500 GDD (60) can be used to monitor crop health and predict yield in a normal weather condition.

Abstmcr-Serif features on reticles are demonstrated to be an effective means to reduce photolithographic pattern errors (such as feature rounding and truncation) caused by the finite resolution of the printing system. In this work, serifs were designed for the thin-oxide level pattern of a CMOS SRAM cell with submicrometer design rules, using two-dimensional aerial image simulations for optimization. Results of experiments using I-line lithography showed goad agreement with the simulations, and significant pattern shape improvement by using serifs was observed. The proximity effects in the electron-beam mask making process were found to presently reduce the full effectiveness of this technique due to the resulting deviations in serif sizes on the reticle. Further enhancement of pattern fidelity may be possible if this issue is resolved.

This paper compares several feature detectors applied to imagery from an unmanned aerial vehicle to find the best detection algorithm when applied to datasets that vary in translation and have little or no image overlap. Metrics of inliers and reconstruction accuracy of feature detectors are considered with respect to three-dimensional reconstruction results. The image matching results are tested experimentally, and an approach to detecting false matches is outlined. Results showed that although the detectors varied in the number of keypoints generated, a large number of inliers does not necessarily translate into more points in the final point cloud reconstruction and that the process of comparing a large quantity of redundant keypoints may outweigh the advantage of having the extra points. The results also showed that despite the development of keypoint detectors and descriptors, none of them consistently demonstrated a substantial improvement in the quality of structure from motion reconstruction when applied to a wide range of disparate urban and rural images.

Nowadays, aerial vehicles (drones) are becoming more popular. Over the past few years, Unmanned Aerial Vehicles (UAVs) have been used in various remote sensing applications. Every aerial vehicle is now either partially or completely automated. The tiniest type of aerial vehicle is the UAV. The widespread use of aerial drones requires numerous safe landing site detection techniques. The paper aims to review literature on techniques for automatic safe landing of aerial drone vehicles by detecting suitable landing sites, considering factors such as ground surfaces and using image processing methods. A drone must determine whether the landing zones are safe for automatic landing. Onboard visual sensors provide potential information on outdoor and indoor ground surfaces through signals or images. The optimal landing locations are then determined from the input data using various image processing and safe landing area detection (SLAD) methods. UAVs are acquisition systems that are quick, efficient, and adaptable. We discuss existing safe landing detection approaches and their achievements. Furthermore, we focus on possible areas for improvement, strength, and future approaches for safe landing site detection. The research addresses the increasing need for safe landing site detection techniques in the widespread use of aerial drones, allowing for automated and secure landing operations

Son yıllarda yüksek konumsal çözünürlüklü uydu görüntülerinin sınıflandırılmasında geleneksel piksel tabanlı sınıflandırma yaklaşımı yerine obje tabanlı yaklaşımın kullanımı önem kazanmıştır. Obje tabanlı görüntü analizi iki temel işlem adımından oluşur. Bunlardan ilki benzer spektral özelliklere sahip piksellerin gruplandırılarak homojen yapılı görüntü objelerinin oluşturulmasını içeren segmentasyon işlemidir. İkincisi ise oluşturulan görüntü objelerin spektral, istatistiksel, geometrik ve ilişkisel özelliklerinden yararlanarak gerçekleştirilen sınıflandırma işlemidir. Sınıflandırma işleminde üretilen görüntü objeleri temel sınıflandırma birimi olarak değerlendirmeye alındığından, obje tabanlı görüntü analizinin doğruluğu segmentasyon işleminin kalitesi ile doğrudan ilişkilidir. Görüntü segmentasyonu ölçek, şekil ve yoğunluk olarak bilinen üç farklı parametre yardımıyla gerçekleştirilmekte olup, bunlar arasında görüntü objelerinin büyüklüğünü kontrol eden ölçek parametresinin tespiti segmentasyonun başarısı açısından büyük bir öneme sahiptir. Bu çalışmada yüksek çözünürlüklü WorldView-2 uydu görüntüsünün obje tabanlı yaklaşımla sınıflandırılması problemi ele alınmıştır. Bu çalışmada, görüntü segmentasyonu işleminin gerçekleştirilmesinde tüm görüntü için tek bir ölçek parametresi uygulamak yerine, görüntüdeki doğal ve yapay nesnelerin dağılımları göz önüne alan bölgesel ölçek parametresi kullanımı önerilmiştir. Tek ölçek parametresi ve farklı ölçek parametreleri ile oluşturulan görüntü objelerinin sınıflandırılmasında en yakın komşuluk (EYK) ve destek vektör makineleri (DVM) algoritmaları kullanılmıştır. Sonuçlar, farklı ölçek parametreleri kullanımıyla DVM ve EYK algoritmalarının hesaplanan sınıflandırma doğruluklarında %6 ve %3’e varan seviyede önemli iyileşmeler olduğunu göstermektedir. McNemar’s testi sonuçlarına göre algoritmaların sınıflandırma performansındaki bu artışların istatistiksel olarak anlamlı olduğu ortaya koyulmuştur

Automated extraction of areas of interest from aerial images is a central issue for map makers. The road skeleton is one of the most frequent targets, and then the prefiltering of road structures is highly important; however, the isolation of man-made structures from a natural landscape is technically tricky because image parts to represent natural geographical features are not uniform patterns. In the present study, we focused on the function of the Morphological Component Analysis (MCA) method to extract structural patterns when dictionaries were appropriately given and demonstrated the effectiveness of edge detection if the prefiltering was done by using MCA. MCA decomposes the image into two patterns in our computer experiment with Curvelet transform and Local Discrete Cosine Transform (LDCT) dictionaries. This approach will explore extensive possibilities of structural data extraction from complex images as an automated method.

Land management, urban planning, and legal compliance all depend on the prompt and accurate detection of
encroachments and the establishment of land boundaries. Especially in vast or complicated land regions, traditional
surveying methods are labor-intensive and prone to inaccuracy since they frequently rely on manual processes or
GPS-based approaches. In this study, we offer a unique method to automate land boundary recognition and monitor
encroachment in satellite and aerial photos using transformer-based deep learning models, specifically Vision
Transformers (ViT). By utilizing transformers’ capacity to identify local and global patterns in high-resolution
images, the suggested approach offers scalability and accuracy improvements over current techniques. Our approach
compares predicted boundaries with historical geospatial data to detect encroachments by integrating boundary
detection with a change detection pipeline. Extensive experiments on publicly available and custom geospatial
datasets demonstrate that the transformer model outperforms traditional convolutional neural networks (CNNs) in
terms of precision, recall, and Intersection over Union (IoU) scores. This paper also explores the integration of the
proposed system with Geographic Information Systems (GIS) for real-time monitoring and visualization. The results
indicate that transformer-based models can revolutionize land surveying by providing a robust, scalable, and efficient
solution for boundary and encroachment detection.

This paper presents a complete high resolution aerial-images processing workflow to detect and characterize vegetation structures in high density urban areas. We present a hierarchical strategy to extract, analyze and delineate vegetation areas according to their height. To detect urban vegetation areas, we develop two methods, one using spectral indices and the second one based on a Support Vector Machines (SVM) classifier. Once vegetation areas detected, we differentiate lawns from treed areas by computing a texture operator on the Digital Surface Model (DSM). A robust region growing method based on the DSM is proposed for an accurate delineation of tree crowns. Delineation results are compared to results obtained by a Random Walk region growing technique for tree crown delineation. We evaluate the accuracy of the tree crown delineation results to a reference manual delineation. Results obtained are discussed and the influential factors are put forward.

An approach for tree species classification in urban areas from high resolution colour infrared (CIR) aerial images and the corresponding Digital Surface Model (DSM) is described in this paper. The proposed method is a supervised classification one based on a Support Vector Machines (SVM) classifier. Texture features from the Gray Level Co-occurrence Matrix (GLCM) are computed to form feature vectors for both per-pixel and per-region classification approaches. The two approaches are presented and results obtained are evaluated and compared both against each other and also against a manual defined ground truth. To perform tree species classification on highdensity urban area images, trees must previously be segmented into individual objects. All intermediary methods developed to segment individual trees will also be shortly described. Tree parameters (height, crown diameter) are estimated from the DSM. These parameters together with the tree species information are used for a 3D realistic modelling of the trees in urban environments. Results of the described system are presented for a typical scene.

Object detection in aerial images has been an active research area thanks to the vast availability of unmanned aerial vehicles (UAVs). Along with the increase of computational power, deep learning algorithms are commonly used for object detection tasks. However, aerial images have large variations, and the object sizes are usually small, rendering lower detection accuracy. Besides, real-time inferencing on low-cost edge devices remains an open-ended question. In this work, we explored the usage of state-of-the-art deep learning object detection on low-cost edge hardware. We propose YOLO-RTUAV, an improved version of YOLOv4-Tiny, as the solution. We benchmarked our proposed models with various state-of-the-art models on the VAID and COWC datasets. Our proposed model can achieve higher mean average precision (mAP) and frames per second (FPS) than other state-of-the-art tiny YOLO models, especially on a low-cost edge device such as the Jetson Nano 2 GB. It was observed that the Jetson Nano 2 GB can achieve up to 12.8 FPS with a model size of only 5.5 MB.

First and foremost, I would like to thank my advisor, Jean-Philippe Avouac, for his continuous support and unconditional enthusiasm. He has inspired me in many ways, and I value the time he has always allocated for our discussions, even on days when his schedule was already unreasonably booked. His approachable attitude has created a friendly and efficient environment, which has encouraged collaborations, and has provided work conditions that everyone could hope for. Pursuing graduate studies in Electrical Engineering while being immersed in the Geological and Planetary Science Division has not always been a simple task. I still remember attending the tecto-lunch group meetings, lost among Earth scientists discussing subjects as weird as rock dating, subduction zones, convergence rates, exhumation rates, etc..., while hoping that someone would just start talking about "normal" topics, such as Fourier analysis. Those were the years when Martine Simoes was trying to teach me tectonics, and when Laurent Demanet was offering me, during our frequent dinners, the stimulating discussions on Fourier analysis I had hoped for earlier during the day. I thank them both to have complementarily impacted my research's views. To defend this thesis, I have gathered professors from several disciplines in Electrical Engineering and in Earth Sciences, and my gratitude goes to Mark Simons, Pietro Perona, P. P. Vaidyanathan, and Charles Elachi, for serving on my committee. I am indebted to Renaud Binet and to Rémi Michel who have introduced me, during my visits to the Laboratoire de Détection et de Géophysique, CEA, France, to their pioneering work on subpixel correlation. The early work with Sylvain Barbot has also proved to be essential, and I am glad that he is now among the regular users of COSI-Corr. The release and the ease of use of the COSI-Corr software would have never been achieved without the work of François Ayoub. His contribution has been determinant in developing the aerial modules as well as the extensions to satellite sensors other than SPOT, such as the ASTER

We describe and test a procedure to accurately co-register and correlate multi-temporal aerial images. We show that this procedure can be used to measure surface deformation, and explore the performance and limitations of the technique. The algorithms were implemented in a software package, COSI-Corr (available from the Caltech Tectonics Observatory website). The technique is validated on several case examples of co-seismic deformation. First, we measure co-seismic ground deformation due to the 1992, Mw 7.3, Landers, California, earthquake from 1 m resolution aerial photography of the National Aerial Photography Program (United States Geological Survey). The fault ruptures are clearly detected, including small kilometric segments with fault slip as small as a few tens of centimeters. We also obtained similar performance from images of the fault ruptures produced by the 1999 Mw 7.1 Hector Mine, California, earthquake. The measurements are shown to be biased due to the inaccuracy of the Digital Elevation Model, film distortions, scanning artifacts, and ignorance of ground displacements at the location of the tie points used to co-register the multi-temporal images. We show that some of these artifacts can be identified and corrected.

Hydromorphology of rivers assessed through direct measurements is a time-consuming and relatively expensive procedure. The rapid development of unmanned aerial vehicles and machine learning (ML) technologies enables the usage of aerial images to determine hydromorphological units (HMUs) automatically. The application of various direct and indirect data sources and their combinations for the determination of river HMUs from aerial images was the main aim of this research. Aerial images with and without the Sobel filter, a layer of boulders identified using Yolov5x6, and a layer of direct measurements of depth and streamflow velocity were used as data sources. Three ML models were constructed for the cases if one, two, or three data sources were used. The ML models for HMU segmentation were constructed of MobileNetV2 pre-trained on ImageNet data for the feature extraction part and U-net for the segmentation part. The stratified K-fold cross-validation with five folds was carried out to evaluate the performance of the model due to the limited dataset. The analysis of the ML results showed that the measured metrics of segmentation using direct measurements were close to the ones of the model trained only on the combination of boulder layer and aerial images with the Sobel filter. The obtained results demonstrated the potential of the applied approach for the determination of HMUs only from the aerial images, and provided a basis for further development to increase its accuracy.

In the last few years, high-resolution imaging of vineyards, obtained by unmanned aerial vehicle recognitions, has provided new opportunities to obtain valuable information for precision farming applications. While available semi-automatic image processing algorithms are now able to detect parcels and extract vine rows from aerial images, the identification of single plant inside the rows is a problem still unaddressed. This study presents a new methodology for the segmentation of vine rows in virtual shapes, each representing a real plant. From the virtual shapes, an extensive set of features is discussed, extracted and coupled to a statistical classifier, to evaluate its performance in missing plant detection within a vineyard parcel. Passing from continuous images to a discrete set of individual plants results in a crucial simplification of the statistical investigation of the problem.

Several high-performing resists identified in the past two years have been exposed at the 0.3-numerical-aperture (NA) SEMATECH Berkeley Microfield Exposure Tool (BMET) with an engineered dipole illumination optimized for 18-nm half pitch. Five chemically amplified platforms were found to support 20-nm dense patterning at a film thickness of approximately 45 nm. At 19-nm half pitch, however, scattered bridging kept all of these resists from cleanly resolving larger areas of dense features. At 18-nm half pitch, none of the resists were are able to cleanly resolve a single line within a bulk pattern. With this same illumination a directly imageable metal oxide hardmask showed excellent performance from 22-nm half pitch to 17-nm half pitch, and good performance at 16-nm half pitch, closely following the predicted aerial image contrast. This indicates that observed limitations of the chemically amplified resists are indeed coming from the resist and not from a shortcoming of the exposure tool. The imageable hardmask was also exposed using a Pseudo Phase-Shift-Mask technique and achieved clean printing of 15-nm half pitch lines and modulation all the way down to the theoretical 12.5-nm resolution limit of the 0.3-NA SEMATECH BMET.

Most publicly available datasets for image classification are with single labels, while images are inherently multilabeled in our daily life. Such an annotation gap makes many pre-trained single-label classification models fail in practical scenarios. This annotation issue is more concerned for aerial images: Aerial data collected from sensors naturally cover a relatively large land area with multiple labels, while annotated aerial datasets, which are publicly available (e.g., UCM, AID), are single-labeled. As manually annotating multi-label aerial images would be time/labor-consuming, we propose a novel selfcorrection integrated domain adaptation (SCIDA) method for automatic multi-label learning. SCIDA is weakly supervised, i.e., automatically learning the multi-label image classification model from using massive, publicly available single-label images. To achieve this goal, we propose a novel Label-Wise self-Correction (LWC) module to better explore underlying label correlations. This module also makes the unsupervised domain adaptation (UDA) from single-to multi-label data possible. For model training, the proposed model only uses single-label information yet requires no prior knowledge of multi-labeled data; and it predicts labels for multi-label aerial images. In our experiments, trained with single-labeled MAI-AID-s and MAI-UCM-s datasets, the proposed model is tested directly on our collected Multi-scene Aerial Image (MAI) dataset. The code and data are available on GitHub().

Photogrammetry with fault fracture density analysis adds high-resolution image of surface structure so that landslide potential of research area can be mapped comprehensively. The main notion in this paper is to distinguish regions and excerpt individual components from an unmanned aerial vehicle (UAV) image comparing to satellite image. Using Chan-Vese segmentation approaches to, we detect regions around the edge around the images. This paper presents an effective initialization method for segmentation using the Chan-Vese model by varying on the iteration. The method is tested on several models and provides considerable rise in performance. Finally, we developed program in Matlab for unravelling resulting equation numerically in fault zone case.

Detection of buildings and other objects from aerial images has various applications in urban planning and map making. Automated building detection from aerial imagery is a challenging task, as it is prone to varying lighting conditions, shadows and occlusions. Convolutional Neural Networks (CNNs) are robust against some of these variations, although they fail to distinguish easy and difficult examples. We train a detection algorithm from RGB-D images to obtain a segmented mask by using the CNN architecture DenseNet. First, we improve the performance of the model by applying a statistical re-sampling technique called Bootstrapping and demonstrate that more informative examples are retained. Second, the proposed method outperforms the non-bootstrapped version by utilizing only one-sixth of the original training data and it obtains a precision-recall break-even of 95.10% on our aerial imagery dataset.

Building extraction from aerial images has several applications in problems such as urban planning, change detection, and disaster management. With the increasing availability of data, Convolutional Neural Networks (CNNs) for semantic segmentation of remote sensing imagery has improved significantly in recent years. However, convolutions operate in local neighborhoods and fail to capture non-local features that are essential in semantic understanding of aerial images. In this work, we propose to improve building segmentation of different sizes by capturing long-range dependencies using contextual pyramid attention (CPA). The pathways process the input at multiple scales efficiently and combine them in a weighted manner, similar to an ensemble model. The proposed method obtains state-of-the-art performance on the Inria Aerial Image Labelling Dataset with minimal computation costs. Our method improves 1.8 points over current state-ofthe-art methods and 12.6 points higher than existing baselines on the Intersection over Union (IoU) metric without any postprocessing. Code and models will be made publicly available.

The data produced by photogrammetric methods have become used more effectively and efficiently to take place of digital aerial cameras to aerial film cameras. Digital aerial cameras have distinct properties according to imaging geometry, sensor numbers, the placement of camera cones and image recording techniques. Since 2000, many studies have been done upon the geometric analysis of the rapidly developing digital imaging system. Intergraph DMC and Vexcel Ultracam are among the most popular digital imaging systems in this area. These cameras are large-format digital aerial camera systems. The purpose of the present study was to investigate the geometric performance of DMCII-250e and Vexcel UltraCamX large-format digital aerial camera systems. The test flights were performed in Bursa province. Two different image blocks which are 10 cm GSD and 30 cm GSD were acquired at different flight altitudes. The image measurements for aerial triangulation were performed by Mescioglu Engineering...

A methodology to estimate water status of palm trees from aerial thermal images was developed. Deficit irrigation of 80% in three drip-irrigated date-palm plots in the northern Dead Sea region was manipulated during the winter of 2007 and 2008. An uncooled thermal camera was used for extensive aerial imaging to detect palm trees and pure-canopy pixels by using only aerial thermal images. An automatic procedure, based on watershed segmentation analysis, was developed which enabled detection of all palm trees in the thermal images. Two new methods were developed to select palm trees and pure pixels within them: basin-based and pixel-based. From the temperatures of pure-canopy pixels, significant differences were found between palm trees under commercial and deficit irrigation regimes, in all three plots. Automated detection of canopy, based on aerial thermal images, is a key step towards commercial mapping of within-plot water-status variability. A protocol, based on the developed methodology, was suggested for mapping water status variability in a palm plot, and for irrigation scheduling.

The problem of distinguishing objects has plagued researchers for many years because of low accuracy compared to human eyes’ capability. In the last decade, the use of Machine Learning in aerial imagery data processing has multiplied, with the technology behind it has also developed exponentially. One of those technologies is image-based object identification, which relies heavily upon data computation. To reduce the computational load, various data segmentation algorithm was developed. This study is focused on reviewing the various image segmentation technology in aerial imagery for image recognition. Literature from as far as 1981 from various journals and conferences worldwide was reviewed. This review examines specific research questions to analyze image segmentation research over time and the challenges researchers face with each method. Machine Learning has gained popularity among segmentation methods. However, Deep Learning has been aggressively put an essential role in it by...

This paper presents a new approach by training and improving a convolutional neural network (CNN) based on You Only Look Once version 2 (YOLOv2) to efficiently detect fuel trucks from images in embedded systems. The proposed method considers the entire image area for strong object detection compared with existing methods that only focus on the image area where the class object exists to predict its probability to be in a class. The loss function for CNN is improved to enhance effective learning, especially when only a limited amount of data is available for training. The class probability can be learned by improving the loss function although the anchor boxes are not in the center of the target object. The learning process of the model can be in a limited range and achieve rapid convergence although the sizes of the initial anchor and target boundary boxes are different. Experimental results of various fuel truck images show the efficiency of the proposed approach under different detection scenarios of real fuel trucks. The detection rate of the proposed method is approximately 4% higher than the YOLOv2 object detection method. The proposed method is suitable to monitor long country borders using unmanned drones.

Purpose Large area traffic monitoring with high spatial and temporal resolution is a challenge that cannot be served by today available static infrastructure. Therefore, we present an automatic near real-time traffic monitoring approach using data of an airborne digital camera system with a frame rate of up to 3 fps. Methods By performing direct georeferencing on the obtained aerial images with the use of GPS/IMU data we are able to conduct near real-time traffic data extraction. The traffic processor consists mainly of three steps which are road extraction supported by a priori knowledge of road axes obtained from a road database, vehicle detection by edge extraction, and vehicle tracking based on normalized cross correlation. Results Traffic data is obtained with a correctness of up to 79% at a completeness of 68%. Conclusions With this system we are able to perform area-wide traffic monitoring with high actuality independent from any stationed infrastructure which makes the syste...

Extreme ultraviolet (EUV) lithography is under development for possible deployment at the 32-nm technology node. One active area of research in this field is the development of photoresists that can meet the stringent requirements (high resolution, high sensitivity, low LER, etc.) of lithography in this regime. In order to facilitate research in this and other areas related to EUV lithography, a printing station based upon the 0.3-NA Micro Exposure Tool (MET) optic was established at the Advanced Light Source, a synchrotron facility at Lawrence Berkeley National Laboratory. A resist modeling technique using a resist point spread function has been shown to have good agreement with experiments for certain EUV resists such as Shipley EUV-2D [2]. The resist point spread function is a two-dimensional function that, when convolved with the simulated aerial image for a given mask pattern and applied to a threshold function, gives a representation of the photoresist pattern remaining after development. The simplicity of this modeling approach makes it attractive for rapid modeling of photoresists for process development applications. In this work, the resist point spread functions for three current high-resolution EUV photoresists [Rohm and Haas EUV-2D, Rohm and Haas MET-1K (XP 3454C), and KRS] are extracted experimentally. This model is then used in combination with aerial image simulations (including effects of projection optic aberrations) to predict the resist pattern for a variety of test patterns. A comparison is made between these predictions and experimental results to evaluate the effectiveness of this modeling technique for newer high-resolution EUV resists.

Line-edge roughness (LER) remains the most significant challenge facing the development of extreme ultraviolet (EUV) resist. The mask, however, has been found to be a significant contributor to image-plane LER. This has long been expected based on modeling and has more recently been demonstrated experimentally. Problems arise from both maskabsorber LER as well as mask multilayer roughness leading to random phase variations in the reflected beam and consequently speckle. Understanding the implications this has on mask requirements for the 22-nm half pitch node and below is crucial. Modeling results indicate a replicated surface roughness (RSR) specification of 50 pm and a ruthenium capping layer roughness specification of 440 pm. Moreover, modeling indicates that it is crucial to achieve the cun'ent ITRS specifications for mask absorber LER which is significantly smaller than current capabilities.

Although extreme ultraviolet (EUV) lithography offers the possibility of very high-resolution patterning, the projection optics must be of extremely high quality in order to meet this potential. One key metric of the projection optic quality is the contrast transfer function (CTF), which is a measure of the aerial image contrast as a function of pitch. A static microfield exposure tool based on the 0.3-NA MET optic and operating at a wavelength of 13.5 nm has been installed at the Advanced Light Source, a synchrotron facility at the Lawrence Berkeley National Laboratory. This tool provides a platform for a wide variety of research into EUV lithography. In this work we present resist-based measurements of the contrast transfer function for the MET optic. These measurements are based upon line/space patterns printed in several different EUV photoresists. The experimental results are compared with the CTF in aerial-image simulations using the aberrations measured in the projection optic using interferometry. In addition, the CTF measurements are conducted for both bright-field and dark-field mask patterns. Finally, the orientation dependence of the CTF is measured in order to evaluate the effect of non-rotationally symmetric lens aberrations. These measurements provide valuable information in interpreting the results of other experiments performed using the MET and similar systems.

Although extreme ultraviolet (EUV) lithography offers the possibility of very high-resolution patterning, the projection optics must be of extremely high quality in order to meet this potential. One key metric of the projection optic quality is the contrast transfer function (CTF), which is a measure of the aerial image contrast as a function of pitch. A static microfield exposure tool based on the 0.3-NA MET optic and operating at a wavelength of 13.5 nm has been installed at the Advanced Light Source, a synchrotron facility at the Lawrence Berkeley National Laboratory. This tool provides a platform for a wide variety of research into EUV lithography. In this work we present resist-based measurements of the contrast transfer function for the MET optic. These measurements are based upon line/space patterns printed in several different EUV photoresists. The experimental results are compared with the CTF in aerial-image simulations using the aberrations measured in the projection optic using interferometry. In addition, the CTF measurements are conducted for both bright-field and dark-field mask patterns. Finally, the orientation dependence of the CTF is measured in order to evaluate the effect of non-rotationally symmetric lens aberrations. These measurements provide valuable information in interpreting the results of other experiments performed using the MET and similar systems.

Although extreme ultraviolet (EUV) lithography offers the possibility of very high-resolution patterning, the projection optics must be of extremely high quality in order to meet this potential. One key metric of the projection optic quality is the contrast transfer function (CTF), which is a measure of the aerial image contrast as a function of pitch. A static microfield exposure tool based on the 0.3-NA MET optic and operating at a wavelength of 13.5 nm has been installed at the Advanced Light Source, a synchrotron facility at the Lawrence Berkeley National Laboratory. This tool provides a platform for a wide variety of research into EUV lithography. In this work we present resist-based measurements of the contrast transfer function for the MET optic. These measurements are based upon line/space patterns printed in several different EUV photoresists. The experimental results are compared with the CTF in aerial-image simulations using the aberrations measured in the projection optic using interferometry. In addition, the CTF measurements are conducted for both bright-field and dark-field mask patterns. Finally, the orientation dependence of the CTF is measured in order to evaluate the effect of non-rotationally symmetric lens aberrations. These measurements provide valuable information in interpreting the results of other experiments performed using the MET and similar systems.

Zhouyuan, the ancient capital of China (B.C1100 to B.C.700), is located at the center of Shaanxi province, covered by farmlands and some modern villages. The continuing archaeological investigations discovered some palace foundations, storage pits and some small tombs, while recent efforts have been focused on locating the foundations of the city walls and the royal tombs. Satellite remote sensing (Landsat TM imagery), aerial photographs and GIS techniques were employed by archaeological research for detection and photogrammetry. Aerial images were used for creating orthomaps, containing information on topography, rivers, roads, modern villages, a.o. to be used as basic background layers for mapping the archaeological sites of the region. Landsat TM images were georeferenced to the orthographic aerial image and combined, they resulted in an image of high ground resolution and wide spectral resolution to be used for interpretation purposes. The above were merged to a GIS, together with an archaeological database containing general information, mainly data and finds of the site, for further analysis. Lasithi (E. Crete) is one of the less-developed and remote districts of the E.U. and at the same time one of the archaeologically most investigated areas in Greece, containing a number of sites from the Neolithic era to the most recent historical period. After a detailed mapping of more than 900 ancient sites (Neolithic-Roman period), research was concentrated on the extraction of the spectral signatures of the sites based on satellite imagery and the cultural resources management of the area. Information regarding the geomorphological and geological attributes of the sites, seismic activity, fires, and climatic conditions were merged into a GIS together with other archaeological information resulting to a definition of risk areas having a variable index of conservation due to the environmetal variables. The above research projects fall within a collaborative framework of research which studies the use of satellite remote sensing and GIS in archaeological research and Cultural Resources Management in semi-arid regions with desertification phenomena, which in turn have direct consequences for the conservation of ancient monuments.

Nowadays, aerial vehicles (drones) are becoming more popular. Over the past few years, Unmanned Aerial Vehicles (UAVs) have been used in various remote sensing applications. Every aerial vehicle is now either partially or completely automated. The tiniest type of aerial vehicle is the UAV. The widespread use of aerial drones requires numerous safe landing site detection techniques. The paper aims to review literature on techniques for automatic safe landing of aerial drone vehicles by detecting suitable landing sites, considering factors such as ground surfaces and using image processing methods. A drone must determine whether the landing zones are safe for automatic landing. Onboard visual sensors provide potential information on outdoor and indoor ground surfaces through signals or images. The optimal landing locations are then determined from the input data using various image processing and safe landing area detection (SLAD) methods. UAVs are acquisition systems that are quick, efficient, and adaptable. We discuss existing safe landing detection approaches and their achievements. Furthermore, we focus on possible areas for improvement, strength, and future approaches for safe landing site detection. The research addresses the increasing need for safe landing site detection techniques in the widespread use of aerial drones, allowing for automated and secure landing operations

Methods for automated orientation of aerial images are presented. They are based on the use of templates, which are derived from existing databases, and area-based matching. The characteristics of available database information and the accuracy requirements for map compilation and orthoimage production are discussed on the example of Denmark. Details on the developed methods for interior and exterior orientation are described. Practical examples like the measurement of réseau images, updating of topographic databases and renewal of orthoimages are used to prove the feasibility of the developed methods.