Animal Detection and Alert System

  This paper presented a farm-deployable animal intrusion system that brings together PIR motion sensing, servo-assisted wide-angle camera coverage, on-device CNN classification, active deterrence, and a resilient dual-channel alert architecture into a single low-cost prototype. The design is shaped specifically by the realities of rural Indian farming: limited budgets, unreliable connectivity, and the practical impossibility of continuous human supervision. Prototype results are encouraging. Overall classification accuracy exceeded 90% for the target species under normal field conditions, and the dual-channel alert pipeline Blynk over Wi-Fi with LoRa as fallback delivered consistent notifications at the distances tested. Active deterrence through light and audio produced reliable retreat behaviour without requiring any intervention from the farmer. Three directions for future development follow directly from observed limitations. First, night-time accuracy degraded to around 78% in very low-light conditions; integrating a low-cost thermal or near-infrared camera module would address this without requiring the floodlight to trigger first. Second, the current model covers five target species; expanding training data to include spotted deer, nilgai, and other regionally prevalent species would widen applicability across more farm contexts. Third, the system currently draws from a 12V DC supply; a solar-charged battery unit with sleep-mode power management would make the design self-sustaining for off-grid deployment a necessary step before large-scale field adoption can be considered.

1.         Delwar, T.S., Mukhopadhyay, S., Kumar, A., Singh, M., Lee, Y., Ryu, J.Y., Hosen, A.S.M.S.: Real-Time Farm Surveillance Using IoT and YOLOv8 for Animal Intrusion Detection. Future Internet 17(2), 70 (2025). 2.        Goyal, N., Sandhu, N.K.: Animal Intrusion Detection and Alert System for Crop Protection. International Journal of Intelligent Systems and Applications in Engineering, 12(10s), 145–151 (2024). 3.        Balakrishna, K., et al.: An IoT-based application with ML integration for animal intrusion detection using R-CNN. IEEE International Conference on Computer Communication and Informatics (ICCCI), Coimbatore (2023). 4.        Alui, et al.: The Prototype of Animal Intrusion Detection and Monitoring Using PIR Sensor and Blynk IoT Platform. Evolution in Electrical and Electronic Engineering, 4(2), 707–715 (2023). 5.        R. R. and Surekha, P.: A Cohesive Farm Monitoring and Wild Animal Warning Prototype System using IoT and Machine Learning. IEEE ICSTCEE, Bengaluru (2020). 6.        Yadav, S., Sharma, N., Yadav, P., et al.: Machine Learning-based Wildlife Intrusion Detector using Raspberry Pi, GSM and CNN. IJSRET, 11(6) (2025). 7.        Zhang, X., et al.: Wildlife Intrusion Detection System using IoT Sensors Integrated with Machine Learning Models for Real-time Classification. IEEE Transactions on Smart Farming (2022). 8.        Chitra, K., et al.: Animals Detection System in the Farm Area Using IoT. IEEE ICCCI, Coimbatore, pp. 1–6 (2023). 9.        Surya, T., Selvi, C.S., Selvaperumal, S.: IoT-based Real-Time Image Processing for Animal Recognition and Classification using Deep Convolutional Neural Network. Microprocessors and Microsystems, 95, 104693 (2022). 10.    Bhanu, K.N., Jasmine, H.J., Mahadevaswamy, H.S.: Machine Learning Implementation in IoT-based Intelligent System for Agriculture. IEEE INCET, Belgaum (2020).