AI-Based Biodiversity Monitoring in Agricultural Landscapes
DOI:
https://doi.org/10.64137/3108088X/IJAES-V2I2P102Keywords:
Artificial Intelligence, Deep Learning, Biodiversity Monitoring, Agroecology, Bioacoustics, Computer Vision, Edge ComputingAbstract
Increased global agricultural intensification threatens biodiversity, ecosystems, and indicator species (ancillary taxa) relevant to the conservation of habitats with harmful effects, including agriculture-driven soil degradation. Conventional methods for monitoring biodiversity are labor-intensive, spatially constrained, and invasive to the monitored organisms inhibiting large-scale and high-frequency conservation actions. In this paper, we propose a novel AI framework tailored for autonomous and non-invasive grazing monitoring of biodiversity within heterogeneous agricultural systems. Based on deep learning-based computer vision for automated entomological and mammalian survey analysis, bioacoustic CNNs (convolutional neural networks) are used to classify avian and amphibian vocalization calls alongside edge-computing deployment. The developed framework provides an entirely automated, scalable technological pathway to ecological health assessment. This AI framework proved to be extremely robust over 12 months of field evaluation in the context of diverse agroecosystems, achieving both high multi-species computer vision identification ($mAP{50}$ =92.4%) and quality isolation from ambient agricultural noise (F1-score=89) quantitative measures for acoustic bird song detection/identification tasks. The use of edge computing also reduced data transmission bandwidth requirements by up to 84.3%, further maximizing the operational viability of remote sensor networks. These findings show that automated AI monitoring can monitor the ecological impacts of sustainable farming with high resolution and at scale, providing an objective window into a data pipeline to reconcile agricultural yield production potential and ecocentric conservation goals.
References
[1] R. Ardila et al., “Common Voice: A Massively-Multilingual Speech Corpus,” arXiv.org, Mar. 05, 2020. https://arxiv.org/abs/1912.06670
[2] J. Jenrette, Z. Y.-C. . Liu, P. Chimote, T. Hastie, E. Fox, and F. Ferretti, “Shark detection and classification with machine learning,” Ecological Informatics, vol. 69, p. 101673, Jul. 2022, doi: https://doi.org/10.1016/j.ecoinf.2022.101673.
[3] S. Christin, É. Hervet, and N. Lecomte, “Applications for deep learning in ecology,” Methods in Ecology and Evolution, Jul. 2019, doi: https://doi.org/10.1111/2041-210x.13256.
[4] G. A. Castellanos‐Galindo and D. R. Robertson, “A tropical fish out of water,” Frontiers in Ecology and the Environment, vol. 18, no. 7, pp. 390–390, Sep. 2020, doi: https://doi.org/10.1002/fee.2247.
[5] J. Redmon and A. Farhadi, “YOLOv3: An Incremental Improvement,” arxiv.org, Apr. 2018, doi: https://doi.org/10.48550/arXiv.1804.02767.
[6] D. Stowell, M. D. Wood, H. Pamuła, Y. Stylianou, and H. Glotin, “Automatic acoustic detection of birds through deep learning: The first Bird Audio Detection challenge,” Methods in Ecology and Evolution, vol. 10, no. 3, pp. 368–380, Nov. 2018, doi: https://doi.org/10.1111/2041-210x.13103.
[7] D. Tuia et al., “Perspectives in machine learning for wildlife conservation,” Nature Communications, vol. 13, no. 1, Feb. 2022, doi: https://doi.org/10.1038/s41467-022-27980-y.
[8] J. Wäldchen and P. Mäder, “Machine learning for image based species identification,” Methods in Ecology and Evolution, vol. 9, no. 11, pp. 2216–2225, Sep. 2018, doi: https://doi.org/10.1111/2041-210x.13075.
[9] S. Xie, R. Girshick, P. Dollár, Z. Tu, and K. He, “Aggregated Residual Transformations for Deep Neural Networks,” IEEE Xplore, Jul. 01, 2017. https://ieeexplore.ieee.org/document/8100117/
[10] Javad Roostaei, Y. Z. Wager, W. Shi, T. M. Dittrich, C. A. Miller, and K. Gopalakrishnan, “IoT-based edge computing (IoTEC) for improved environmental monitoring,” Sustainable Computing: Informatics and Systems, vol. 38, pp. 100870–100870, Apr. 2023, doi: https://doi.org/10.1016/j.suscom.2023.10087z0.
