Animal Migration Patterns Extraction Based on Atrous-Gated CNN Deep Learning Model
Abstract
:1. Introduction
2. Materials
2.1. Scanning Strategy and Data Products
2.2. Data Rendering
- Significant differences exist in the reflectivity of biological and precipitation echo. Biological targets have smaller reflection area than precipitation, which results in echo intensity much lower than precipitation echo. The intensity of precipitation echo can exceed 50 dBZ, but usually lower than 20 dBZ for biological echo [11,28].
- Biological targets and precipitation targets have different airspeeds. Doppler velocity parameters are usually used to distinguish different types of target scatterers (birds, bats, insects, raindrops, and other particles) [11,29]. Most birds fly at speeds greater than 8 m/s, while air speeds are generally lower [30]. Daytime fain weather without birds produce weather signals that have the smallest volumetric median values of spectral width (i.e., <2 m/s), echoes contaminated by birds usually have a higher spectral width (i.e., >12 m/s) [11,31,32].
2.3. Data Annotations
2.3.1. Precipitation Echo
- Primarily visible in the red and green channel. Precipitation echo can extend to several kilometers in the vertical direction [34] while being scanned by weather radar at all elevation angles, and the echo intensity can exceed 50 dBZ [11,28]; thus, the red and green channels of the rendered image have strong gray values. In addition, in environments without biological echo pollution, including isolated tornado storms, the echo spectrum width is usually lower than 2 m/s [31,32], meaning the third channel tends to have a black background.
- Echo shape is irregular especially at the edge, and the texture is coarser. Precipitation echo is greatly affected by the underlying surface and can change widely within the radar detection area [35], resulting in more irregular shapes and textures.
- Significant displacement changes in time and space.
2.3.2. Biological Echo
- Primarily visible in the red and blue channel. The migration altitude of birds usually occurs below 3000 m [11] and can be almost completely covered by weather radar’s 0.5° elevation beam at 200 km. Higher elevation angles cannot effectively illuminate biological echoes in a large-scale range [26,35]. In addition, the migration of biological groups, especially when migrating at different radial speeds, will cause a higher spectral width [31,32].
- Echo shape is approximately circular, and the texture and edges are smoother. Animal migration is usually concentrated at a certain height, with obvious layered characteristics [36], meaning biological populations are more likely evenly distributed relative to precipitation [35], and the maximum migration height remains basically unchanged. As a result, under the observation of polar coordinates, the farthest detection range of different azimuth angles remains basically unchanged (Figure 2), thus the echo is approximately circular in rectangular coordinate.
- Fast gathering in time and space. Large-scale migration of aerial creatures usually occurs during the evening [11], causing a sudden increase in the number of scattered targets in the airspace detected by the weather radar, which will result in a sudden increase in the intensity of the echo.
3. Methods
- We incorporate Squeeze-and-Excitation (SE) [38] block into the regular stream. For our rendered radar images, the channel R and the channel G corresponds to reflectivity image with elevation angle of 0.5° and 1.45°, respectively. The channel R is mainly dominated by biological migration events, while the channel G is mainly dominated by precipitation events. We suppose that the two events are independent and SE block can reduce channel interdependence as well as enhance the attention on the current feature map.
- We add atrous spatial pyramid pooling (ASPP) [39] block to build Atrous Gated Convolution (AGC) before Gated Convolutional Layers [37] in the original network. For our rendered radar images, it is normal that biological and precipitation targets occupy large-scale spatial areas, resulting in wide coverage in PPI scans. ASPP block can expand the receptive field, helping the network understand the global shape information.
3.1. Gated-SCNN
3.2. SE Block
3.3. ASPP Block
4. Results
4.1. Dataset
4.2. Evaluation and Experimental Settings
4.3. CNN Performance
4.3.1. Quantitative Evaluation
4.3.2. Qualitative Evaluation
4.3.3. Computational Complexity
5. Discussion
6. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
References
- Hahn, S.; Bauer, S.; Liechti, F. The natural link between Europe and Africa–2.1 billion birds on migration. Oikos 2009, 118, 624–626. [Google Scholar] [CrossRef]
- Dokter, A.M.; Farnsworth, A.; Fin, D. Seasonal abundance and survival of North America’s migratory avifauna determined by weather radar. Nat. Ecol. Evol. 2018, 2, 1603–1609. [Google Scholar] [CrossRef] [PubMed]
- Van, D.B.M.; Horton, K.G. A continental system for forecasting bird migration. Science 2018, 361, 1115–1118. [Google Scholar] [CrossRef]
- Tsiodras, S.; Kelesidis, T.; Kelesidis, I. Human infections associated with wild birds. J. Infect. 2008, 56, 83–98. [Google Scholar] [CrossRef]
- Zakrajsek, E.J.; Bissonette, J.A. Ranking the risk of wildlife species hazardous to military aircraft. Wildl. Soc. Bull. 2005, 33, 258–264. [Google Scholar] [CrossRef]
- Haykin, S.; Deng, C. Classification of radar clutter using neural networks. IEEE Trans. Neural Netw. 1991, 2, 589–600. [Google Scholar] [CrossRef]
- Brooks, M. Electronics as a possible aid in the study of bird flight and migration. Science 1945, 101, 329. [Google Scholar] [CrossRef] [PubMed]
- Lack, D.; Varley, G.C. Detection of birds by radar. Nature 1945, 156, 446. [Google Scholar] [CrossRef]
- Gauthreaux, J.S.A.; Belser, C.G. Radar ornithology and biological conservation. AUK 2003, 120, 266–277. [Google Scholar] [CrossRef]
- Li, W.; Hu, C.; Wang, R. Comprehensive analysis of polarimetric radar cross-section parameters for insect body width and length estimation. Sci. China Inf. Sci. 2021, 64, 122302. [Google Scholar] [CrossRef]
- Gauthreaux, S.A.; Belser, C.G. Displays of bird movements on the WSR-88D: Patterns and quantification. Weather Forecast. 1998, 13, 453–464. [Google Scholar] [CrossRef]
- Stepanian, P.M.; Wainwright, C.E. Ongoing changes in migration phenology and winter residency at Bracken Bat Cave. Glob. Change Biol. 2018, 24, 3266–3275. [Google Scholar] [CrossRef] [PubMed]
- Rennie, S.J. Common orientation and layering of migrating insects in southeastern Australia observed with a Doppler weather radar. Meteorol. Appl. 2014, 21, 218–229. [Google Scholar] [CrossRef]
- Lin, T.Y.; Winner, K.; Bernstein, G. MistNet: Measuring historical bird migration in the US using archived weather radar data and convolutional neural networks. Methods Ecol. Evol. 2019, 10, 1908–1922. [Google Scholar] [CrossRef]
- Schmaljohann, H.; Liechti, F.; Bächler, E. Quantification of bird migration by radar—A detection probability problem. Ibis 2008, 150, 342–355. [Google Scholar] [CrossRef]
- Dokter, A.M.; Desmet, P.; Spaaks, J.H. bioRad: Biological analysis and visualization of weather radar data. Ecography 2019, 42, 852–860. [Google Scholar] [CrossRef] [Green Version]
- Gauthreaux, S.; Diehl, R. Discrimination of biological scatterers in polarimetric weather radar data: Opportunities and challenges. Remote Sens. 2020, 12, 545. [Google Scholar] [CrossRef] [Green Version]
- Hu, C.; Li, S.; Wang, R. Extracting animal migration pattern from weather radar observation based on deep convolutional neural networks. J. Eng. Technol. 2019, 2019, 6541–6545. [Google Scholar] [CrossRef]
- Gauthreaux, S.A.; Livingston, J.W.; Belser, C.G. Detection and discrimination of fauna in the aerosphere using Doppler weather surveillance radar. Integr. Comp. Biol. 2008, 48, 12–23. [Google Scholar] [CrossRef] [Green Version]
- O’Neal, B.J.; Stafford, J.D.; Larkin, R.P. Waterfowl on weather radar: Applying ground-truth to classify and quantify bird movements. J. Field Ornithol. 2010, 81, 71–82. [Google Scholar] [CrossRef]
- Lakshmanan, V.; Fritz, A.; Smith, T. An automated technique to quality control radar reflectivity data. J. Appl. Meteorol. Climatol. 2007, 46, 288–305. [Google Scholar] [CrossRef] [Green Version]
- Lakshmanan, V.; Zhang, J.; Howard, K. A technique to censor biological echoes in radar reflectivity data. J. Appl. Meteorol. Climatol. 2010, 49, 453–462. [Google Scholar] [CrossRef] [Green Version]
- Dokter, A.M.; Liechti, F.; Stark, H. Bird migration flight altitudes studied by a network of operational weather radars. J. R. Soc. Interface 2011, 8, 30–43. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- RoyChowdhury, A.; Sheldon, D.; Maji, S. Distinguishing weather phenomena from bird migration patterns in radar imagery. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, Las Vegas, NV, USA, 19 December 2016; pp. 10–17. [Google Scholar]
- Stepanian, P.M.; Horton, K.G.; Melnikov, V.M. Dual-polarization radar products for biological applications. Ecosphere 2016, 7, e01539. [Google Scholar] [CrossRef]
- Cui, K.; Hu, C.; Wang, R. Deep-learning-based extraction of the animal migration patterns from weather radar images. Sci. China Inf. Sci. 2020, 63, 140304. [Google Scholar] [CrossRef] [Green Version]
- Larkin, R.P.; Diehl, R.H. Spectrum width of flying animals on Doppler radar. In Proceedings of the 15th Conference on Biometeorology/Aerobiology and 16th International Congress of Biometeorology, American Meteorological Society, Kansas City, MI, USA, 30 October 2002; p. 9. [Google Scholar]
- Rennie, S.J.; Curtis, M.; Peter, J. Bayesian echo classification for Australian single-polarization weather radar with application to assimilation of radial velocity observations. J. Atmos. Ocean. Technol. 2015, 32, 1341–1355. [Google Scholar] [CrossRef]
- Gauthreaux, S.A.; Belser, C.G. Bird movements on Doppler weather surveillance radar. Birding 2003, 35, 616–628. [Google Scholar]
- Bruderer, B.; Boldt, A. Flight characteristics of birds: I. Radar measurements of speeds. Ibis 2001, 143, 178–204. [Google Scholar] [CrossRef]
- Feng, H.; Wu, K.; Cheng, D. Spring migration and summer dispersal of Loxostege sticticalis (Lepidoptera: Pyralidae) and other insects observed with radar in northern China. Environ. Entomol. 2004, 33, 1253–1265. [Google Scholar] [CrossRef]
- Fang, M.; Doviak, R.J.; Melnikov, V. Spectrum width measured by WSR-88D: Error sources and statistics of various weather phenomena. J. Atmos. Ocean. Technol. 2004, 21, 888–904. [Google Scholar] [CrossRef]
- Rosa, I.M.; Marques, A.T.; Palminha, G. Classification success of six machine learning algorithms in radar ornithology. Ibis 2016, 158, 28–42. [Google Scholar] [CrossRef]
- Lamkin, W.E. Radar Signature Analysis of Weather Phenomena. Ann. N. Y. Acad. Sci. 1969, 163, 171–186. [Google Scholar] [CrossRef]
- Zhang, P.; Liu, S.; Xu, Q. Identifying Doppler velocity contamination caused by migrating birds. Part I: Feature extraction and quantification. J. Atmos. Ocean. Technol. 2005, 22, 1105–1113. [Google Scholar] [CrossRef]
- Hu, G.; Lim, K.S.; Horvitz, N. Mass seasonal bioflows of high-flying insect migrants. Science 2016, 354, 1584–1587. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Takikawa, T.; Acuna, D.; Jampani, V. Gated-scnn: Gated shape cnns for semantic segmentation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, Seoul, Korea, 27 October–2 November 2019; pp. 5229–5238. [Google Scholar]
- Hu, J.; Shen, L.; Sun, G. Squeeze-and-excitation networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA, 18–23 June 2018; pp. 7132–7141. [Google Scholar] [CrossRef] [Green Version]
- Chen, L.C.; Papandreou, G.; Kokkinos, I. Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs. IEEE Trans. Pattern Anal. Mach. Intell. 2017, 40, 834–848. [Google Scholar] [CrossRef]
- Yu, F.; Koltun, V. Multi-scale context aggregation by dilated convolutions. arXiv 2016, arXiv:1511.07122v3. [Google Scholar]
Number | Parameter | Value |
---|---|---|
1 | Range Bin Number of Reflectivity | 460 |
2 | Range Bin Number of Velocity | 920 |
3 | Range Bin Number of Spectral Width | 920 |
4 | Range resolution of Reflectivity | 1000 m |
5 | Range resolution of Velocity | 250 m |
6 | Range resolution of Spectral Width | 250 m |
7 | Elevation | 0.5°–19.5° |
8 | Beamwidth | 1° |
Echo Target | Gated-SCNN | Atrous-Gated CNN | ||||||
---|---|---|---|---|---|---|---|---|
Precision | Recall | F-Score | IoU | Precision | Recall | F-Score | IoU | |
Background | 0.991 | 0.993 | 0.992 | 0.984 | 0.982 | 0.991 | 0.986 | 0.974 |
Precipitation | 0.983 | 0.974 | 0.978 | 0.958 | 0.992 | 0.981 | 0.985 | 0.974 |
Biology | 0.982 | 0.985 | 0.987 | 0.968 | 0.996 | 0.989 | 0.992 | 0.985 |
Method | Precision | Recall | F-Score | IoU |
---|---|---|---|---|
MistNet | 0.935 | 0.990 | 0.962 | 0.926 |
DeepLab V3+ | 0.985 | 0.948 | 0.968 | 0.937 |
Gated-SCNN | 0.982 | 0.985 | 0.987 | 0.968 |
Gated-SCNN + SE | 0.993 | 0.986 | 0.989 | 0.980 |
Gated-SCNN + ASPP | 0.995 | 0.987 | 0.991 | 0.982 |
Atrous-Gated CNN | 0.996 | 0.989 | 0.992 | 0.985 |
Method | Params (M) | FLOPs (G) |
---|---|---|
MistNet | 125.10 | 93.50 |
DeepLab V3+ | 59.34 | 34.70 |
Gated-SCNN | 136.74 | 291.75 |
Atrous-Gated CNN | 137.32 | 298.30 |
Method | Precision | Recall | F-Score | IoU |
---|---|---|---|---|
MistNet | 0.726 | 0.961 | 0.827 | 0.705 |
DeepLab V3+ | 0.979 | 0.902 | 0.939 | 0.885 |
Gated-SCNN | 0.974 | 0.960 | 0.967 | 0.936 |
Atrous-Gated CNN | 0.983 | 0.966 | 0.974 | 0.950 |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Wang, S.; Hu, C.; Cui, K.; Wang, R.; Mao, H.; Wu, D. Animal Migration Patterns Extraction Based on Atrous-Gated CNN Deep Learning Model. Remote Sens. 2021, 13, 4998. https://doi.org/10.3390/rs13244998
Wang S, Hu C, Cui K, Wang R, Mao H, Wu D. Animal Migration Patterns Extraction Based on Atrous-Gated CNN Deep Learning Model. Remote Sensing. 2021; 13(24):4998. https://doi.org/10.3390/rs13244998
Chicago/Turabian StyleWang, Shuaihang, Cheng Hu, Kai Cui, Rui Wang, Huafeng Mao, and Dongli Wu. 2021. "Animal Migration Patterns Extraction Based on Atrous-Gated CNN Deep Learning Model" Remote Sensing 13, no. 24: 4998. https://doi.org/10.3390/rs13244998