Academia.eduAcademia.edu

Figure 24 – uploaded by Cedrick Rochet

See full PDFdownloadDownload figure
Figure 24: Error rate of 11 classes with different Gaussian mixture function number.

Figure 24 Error rate of 11 classes with different Gaussian mixture function number.

Related Figures (39)

Figure 1: Proposed automated bird monitoring and recognition system.
Figure 1: Proposed automated bird monitoring and recognition system.
FicureE 3: Block diagram of the DOA estimation of a wideband source [5]. Ficure 2: The proposed concentric circular microphone array.
FicureE 3: Block diagram of the DOA estimation of a wideband source [5]. Ficure 2: The proposed concentric circular microphone array.
Ficure 4: Narrowband MUSIC spectrum for 2 DOAs (500 Hz nar- rowband signal).
Ficure 4: Narrowband MUSIC spectrum for 2 DOAs (500 Hz nar- rowband signal).
Figure 5: Narrowband MUSIC spectrum with small peaks from noise removed.
Figure 5: Narrowband MUSIC spectrum with small peaks from noise removed.
Ficure 6: Narrowband MUSIC spectrum with only local maxima and minima.
Ficure 6: Narrowband MUSIC spectrum with only local maxima and minima.
Ficure 7: Narrowband MUSIC spectrum with only local maxima.
Ficure 7: Narrowband MUSIC spectrum with only local maxima.
Ficure 8: Combined narrowband DOA estimates.
Ficure 8: Combined narrowband DOA estimates.
TABLE 1: Statistical performance of the proposed wideband DOA estimation technique for a given DOA. TABLE 2: Statistical performance of the proposed wideband DOA estimation technique for a given DOA.
TABLE 1: Statistical performance of the proposed wideband DOA estimation technique for a given DOA. TABLE 2: Statistical performance of the proposed wideband DOA estimation technique for a given DOA.
Ficure 10: MUSIC spectrum for a linear array.
Ficure 10: MUSIC spectrum for a linear array.
Figure 11: The beam pattern of the proposed circular array at 1 kHz.
Figure 11: The beam pattern of the proposed circular array at 1 kHz.
FiGuRE 12: The proposed circular array configuration. The above discussion is for a narrowband input. For a wideband input, we first separate the incoming data into frames, and apply FFT to decompose the input signal into narrowband components. The above design procedure is then applied at different narrowband components and the resultant output is obtained through inverse FFT. The pro- posed design method above can also achieve frequency in- variant beam pattern and the details can be found in [7].
FiGuRE 12: The proposed circular array configuration. The above discussion is for a narrowband input. For a wideband input, we first separate the incoming data into frames, and apply FFT to decompose the input signal into narrowband components. The above design procedure is then applied at different narrowband components and the resultant output is obtained through inverse FFT. The pro- posed design method above can also achieve frequency in- variant beam pattern and the details can be found in [7].
FiGuRE 13: Grouping of the rings in the four subarrays. TABLE 3: Grouping of rings into different subarrays for broadband beamforming. therefore resulting in savings in array elements. The pro- posed compound ring structure has 4 operating frequency bands as listed in the second column of Table 3. The third column in the table shows the number of rings in each band and the fourth column is the number of elements in each ring for the frequency band considered. The grouping of the rings for different bands is shown in Figure 13. The minimum sep- aration between two rings is
FiGuRE 13: Grouping of the rings in the four subarrays. TABLE 3: Grouping of rings into different subarrays for broadband beamforming. therefore resulting in savings in array elements. The pro- posed compound ring structure has 4 operating frequency bands as listed in the second column of Table 3. The third column in the table shows the number of rings in each band and the fourth column is the number of elements in each ring for the frequency band considered. The grouping of the rings for different bands is shown in Figure 13. The minimum sep- aration between two rings is
Ficure 15: 3D beam pattern of compound linear array (0) = 7/4, do = m/4).
Ficure 15: 3D beam pattern of compound linear array (0) = 7/4, do = m/4).
Figure 16: 3D beam pattern for ring array. (0) = 7/4, do = 7/4).
Figure 16: 3D beam pattern for ring array. (0) = 7/4, do = 7/4).
Ficure 18: Received signal in one channel. Interference signal is coming from a DOA in the ambiguity set. Ficure 17: Beampattern for a linear array.
Ficure 18: Received signal in one channel. Interference signal is coming from a DOA in the ambiguity set. Ficure 17: Beampattern for a linear array.
Figure 19: Output of the linear array. Interference signal is coming from a DOA in the ambiguity set.
Figure 19: Output of the linear array. Interference signal is coming from a DOA in the ambiguity set.
Figure 20: Output of the ring array. Interference signal is coming from a DOA in the ambiguity set.
Figure 20: Output of the ring array. Interference signal is coming from a DOA in the ambiguity set.
Ficure 21: Output of ring array with null at the DOA of 1200 Hz interference. Interference signal is coming from a DOA in the ambi- guity set. A null is created in the direction of the interference signal.
Ficure 21: Output of ring array with null at the DOA of 1200 Hz interference. Interference signal is coming from a DOA in the ambi- guity set. A null is created in the direction of the interference signal.
FiGure 22: Preprocessing steps in the feature extraction subsystem.
FiGure 22: Preprocessing steps in the feature extraction subsystem.
TABLE 4: Summary of bird sound files and feature numbers for each bird class.
TABLE 4: Summary of bird sound files and feature numbers for each bird class.
Ficure 23: Error rate of 11 classes with different feature window length.
Ficure 23: Error rate of 11 classes with different feature window length.
Ficure 26: Test result of using half randomly selected files for train- ing, and the rest half files for testing (20 runs). Ficure 25: Test result of using half features randomly selected from each file for training, and the rest half for testing (30 runs).
Ficure 26: Test result of using half randomly selected files for train- ing, and the rest half files for testing (20 runs). Ficure 25: Test result of using half features randomly selected from each file for training, and the rest half for testing (30 runs).
Figure 28: Circular array configuration.
Figure 28: Circular array configuration.
TABLE 5: Summary of ring radii.
TABLE 5: Summary of ring radii.
TaBLe 6: Angular distribution of array elements. 7.2. Beamformer outputs beamforming. It can be seen that before the beamform- ing, the signals were very noisy. However, after the beam- forming, the signals were very clean. We could not hear any interference and background noise after the beamform- ing.
TaBLe 6: Angular distribution of array elements. 7.2. Beamformer outputs beamforming. It can be seen that before the beamform- ing, the signals were very noisy. However, after the beam- forming, the signals were very clean. We could not hear any interference and background noise after the beamform- ing.
Ficure 30: Finished circular microphone array. It has a dimension of 1.2 meter in diameter.
Ficure 30: Finished circular microphone array. It has a dimension of 1.2 meter in diameter.
TABLE 7: Summary of DOA estimation and the expected DOA val ues. 7.3. Bird classification performance
TABLE 7: Summary of DOA estimation and the expected DOA val ues. 7.3. Bird classification performance
FicurE 32: Signals before beamforming and after beamforming.
FicurE 32: Signals before beamforming and after beamforming.
Y. Zhang received her B.S. and M.S. degrees in automatic control from Shanghai Jiao- tong University in China in 1996 and 1999. She then obtained her M.S. and Ph.D. de- grees in electrical engineering from Geor- gia Institute of Technology in 2001 and 2005, respectively. Since July 2005, she has been with Intelligent Automation, Inc (IAD), USA, where she is currently a Research En- gineer. Her research interests include speech and image processing, intelligent systems, and controls.
Y. Zhang received her B.S. and M.S. degrees in automatic control from Shanghai Jiao- tong University in China in 1996 and 1999. She then obtained her M.S. and Ph.D. de- grees in electrical engineering from Geor- gia Institute of Technology in 2001 and 2005, respectively. Since July 2005, she has been with Intelligent Automation, Inc (IAD), USA, where she is currently a Research En- gineer. Her research interests include speech and image processing, intelligent systems, and controls.
Related topics:
Computer ScienceArtificial IntelligenceMicrophone ArraysSensor ArrayBird Migration MonitoringElectrical and Electronic EngineeringMicrophone array
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved