Optical flow using color information: preliminary results

2009

Abstract—Motion detection is very important in image processing. One way of detecting motion is using optical flow. Optical flow cannot be computed locally, since only one independent measurement is available from the image sequence at a point, while the flow velocity has two components. A second constraint is needed. The method used for finding the optical flow in this project is assuming that the apparent velocity of the brightness pattern varies smoothly almost everywhere in the image.

1995

time-varying 2D projection. Hence as an intermediate The use of color images for motion estimation is investigated stage in the 3D motion estimation process we need to in this work. Beyond the straightforward approach of using find the 2D projection of a motion field onto the image the color components as separate images of the same scene, a plane. This 2D field of velocities is called image flow. new method, based on exploiting color invariance under motion, Then we could attempt to recover the 3D motion field is discussed. Two different sets of color-related, locally computfrom its 2D projection. To do this, however, seems able motion ''invariants'' are analyzed and tested in this paper, impossible without additional information about the and the results of motion estimation based on them are commotion field. Instead, we can extract the so-called pared to the direct use of the RGB brightness functions. © 1997 optical flow, which is a 2D field of velocities associated Academic Press with the variation of brightness patterns of the image [7, 17].

Artificial intelligence in perspective, 1994

Optical flow cannot be computed locally, since only one independent measurement is available from the image sequence at a point, while the flow velocity has two components. A second constraint is needed. A method for finding the optical flow pattern is presented which assumes that the apparent velocity of the brightness pattern varies smoothly almost everywhere in the image. An iterative implementation is shown which successfully computes the optical flow for a number of synthetic image sequences. The algorithm is robust in that it can handle image sequences that are quantized rather coarsely in space and time. It is also insensitive to quantization of brightness levels and additive noise. Examples are included where the assumption of smoothness is violated at singular points or along lines in the image.

2012

This paper presents optical flow estimation technique to estimate the motion vectors in each frame of the video sequence. By thresholding and performing morphological closing on the motion vectors, we produces binary feature images. Using these binary features the cars are located. A bounding Box is drawn around the cars that pass beneath the white line. The algorithm used for this is lucas kanade. Use of the threshold to reduce the noise in small movements between frames is analyzed. Higher the threshold ,the less small movements impact the optical flow calculation. Experiments are done to find the value that best achieves our results.

J. Inf. Sci. Eng., 2008

In this paper, we develop two methods for estimating optical flow across multiframes in an image sequence. The rationale of our proposed methods is based on integrating temporal information from the reference image and its previous and next several images. The difference between the two methods lies mainly on the modeling in utilizing multiple images. The proposed methods can be applied to each point independently, and hence is suitable for many image-based applications. Experiments were conducted on various image sequences and the results show that using multiple frames can indeed reduce the estimation error significantly, especially at those points that possess good features.

2012

Optical flow estimation is often understood to be identical to dense image based motion estimation . However, only under certain assumptions does optical flow coincide with the projection of the actual 3D motion to the image plane . Most prominently, transparent and glossy scene-surfaces or changes in illumination introduce a difference between the motion of objects in the world and the apparent motion. Unfortunately, in the real world, glossy and (semi-) transparent objects as well as changes in illumination, e.g., cast shadows, are rather frequent.

Optical Engineering, 2015

Optical flow (OF) methods are used to estimate dense motion information between consecutive frames in image sequences. In addition to the specific OF estimation method itself, the quality of the input image sequence is of crucial importance to the quality of the resulting flow estimates. For instance, lack of texture in image frames caused by saturation of the camera sensor during exposure can significantly deteriorate the performance. An approach to avoid this negative effect is to use different camera settings when capturing the individual frames. We provide a framework for OF estimation on such sequences that contain differently exposed frames. Information from multiple frames are combined into a total cost functional such that the lack of an active data term for saturated image areas is avoided. Experimental results demonstrate that using alternate camera settings to capture the full dynamic range of an underlying scene can clearly improve the quality of flow estimates. When saturation of image data is significant, the proposed methods show superior performance in terms of lower endpoint errors of the flow vectors compared to a set of baseline methods. Furthermore, we provide some qualitative examples of how and when our method should be used.

This paper builds a novel extracting method by sequentially read frame by frame images or video feed without consuming all of the resource of the equipment. This method calls a function between temporal change of brightness at a target pixel. This one can define the edge of each frame of the image at the local velocity. So we will compare the two images and take the absolute difference between them. Thus it will leave the perfect threshold to represent which pixels change between the two images after comparison then we may estimate analytically from lags times of the correlation functions. We will start by the principle that human eye can see a moving object from the background that is still, however, the object is bouncing or moving because it's related to the background. Sequential Image algorithm can be further refined to yield time complexity. Its result can show the linear speedup and can be achieved quite faster without consuming all system resources.

2011

Abstract Tracking objects in real-time has a variety of applications in many fields. Optical flow based tracking is one such tracking mechanism which can track moving objects even under complex backgrounds and different light conditions. The research presented in this paper discusses the feasibility of using optical flow to track moving objects captured in a camera view, to extract basic information related to the objects.

2009

The quantitative evaluation of optical flow algorithms by led to significant advances in performance. The challenges for optical flow algorithms today go beyond the datasets and evaluation methods proposed in that paper. Instead, they center on problems associated with complex natural scenes, including nonrigid motion, real sensor noise, and motion discontinuities. We propose a new set of benchmarks and evaluation methods for the next generation of optical flow algorithms. To that end, we contribute four types of data to test different aspects of optical flow algorithms: (1) sequences with nonrigid motion where the ground-truth flow is determined by tracking hidden fluorescent texture, (2) realistic synthetic sequences, (3) high frame-rate video used to study interpolation error, and (4) modified stereo sequences of static scenes. In addition to the average angular error used by Barron et al., we compute the absolute flow endpoint error, measures for frame interpolation error, improved statistics, and results at motion discontinuities and in textureless regions. In October 2007, we published the performance of several well-known methods on a preliminary version of our data to establish the current state of the art. We also made the data freely available on the web at http://vision.middlebury.edu/flow/. Subsequently a number of researchers have uploaded their results to our website and published papers using the data. A significant improvement in performance has already been achieved. In this paper we analyze the results obtained to date and draw a large number of conclusions from them.