This paper presents the design and implementation of a Smart Agriculture Wildlife Intrusion Detection and Repellent System using Machine Leaming The system integrates an OV7670 night-vision USB camera for continuous farm monitoring, a YOLOVS deep learning model for real-time animal detection and classification with approximately 94% accuracy, and an ATmega328P microcontroller as the central control unit. Upon confirmed animal detection, a high-intensity LED strobe light and a 5V piezoelectric sound buzzer activate automatically for five seconds as non-harmful repellents A PIR sensor (HC-SR501) enables event-driven system wake-up, reducing energy consumption by approximately 60% compared to alwarys-on systems An automated email alert using Python smtplib dispatches the captured intrusion image to the farmer within two seconds. Experimental results confirm reliable detection of elephants, wild boars, and deer under real night time farm conditions with a complete detection-to-alert pipeline executing in under two seconds

The Smart Agriculture Wildlife Intrusion Detection and Repellent System Using Machine Learning was successfully implemented using an OV7670 night-vision camera, YOLOv5s model, ATmega328P microcontroller (PIR on A0, LED on pin 8, Buzzer on pin 9, LCD on pins 2-7), BC547 transistor drivers, and Python smtplib email alert system. The program runs on a 300ms polling loop with boolean flags to prevent repeated triggering within the same detection event. All four test scenarios produced correct LCD output, repellent response, and email delivery as verified during prototype testing. The system achieved approximately 94% animal detection accuracy with a complete detection-to-alert pipeline executing in under two seconds. The PIR-triggered sleep architecture reduced energy consumption by approximately 60% compared to always-on systems, and the 12V/7Ah battery backup sustained eight hours of uninterrupted night operation. Future work will incorporate edge deployment of YOLOv5 on NVIDIA Jetson Nano to eliminate laptop dependency, GSM-based SMS alerts for internet-unavailable rural areas, solar power integration for fully off-grid sustainable operation, and a cloud-based web dashboard for forest officials to monitor intrusion patterns across multiple farm locations.

[1]Y. K. S. Kiat et al., “Wildlife Intrusion Detection System for Agriculture Protection Based on Image Recognition,” 2024 IEEE 13th Global Conference on Consumer Electronics (GCCE), Kitakyushu, Japan, pp. 366-370, doi: 10.1109/GCCE62371.2024.10760559. [2]M. Kathir, V. Balaji and K. Ashwini, “Animal Intrusion Detection Using YOLOv8,” 2024 10th International Conference on Advanced Computing and Communication Systems (ICACCS), Coimbatore, India, pp. 206-211, doi: 10.1109/ICACCS60874.2024.10716895. [3]Aibin Abraham, Bibin Mathew, Devika Panikkar and Jaya John, “Wild Animal Intrusion Detection System using YOLO,” International Journal of Innovative Science and Research Technology, Vol. 8, Issue 5, May 2023, ISSN: 2456-2165. [4]Sathesh, K. Vishnu et al., “Image Processing based Protection of Crops from Wild Animals using Intelligent Surveillance,” International Conference on Electronics and Renewable Systems, 2022. [5]D. Ranparia et al., “Machine Learning-based Acoustic Repellent System for Protecting Crops against Wild Animal Attacks,” IEEE (ICIIS), 2021. [6] Mohit Korche et al., “Smart Crop Protection System,” International Journal of Latest Engineering Science (IJLES), 2021. [7]K. Mohana Lakshmi et al., “Security for Protecting Agricultural Crops from Wild Animals using GSM Technology,” Journal of Shanghai Jiaotong University, 2020. [8]Ashwini V. Sayagavi, Sudarshan T S B and Prashanth C. Ravoor, “Deep Learning Methods for Animal Recognition and Tracking to Detect Intrusions,” 2020. [9]S. Shaik et al., “Real-Time AI-Based Wildlife Detection and Deterrent System for Farmland Protection,” IEEE Conference on Smart Agriculture Systems, 2024.

© 2025 International Journal of Engineering and Techniques (IJET).

Digital Object Identifier (DOI)

IJET assigns a unique DOI to every accepted article via Zenodo for persistent linking and citation. Your article’s DOI: https://doi.org/{{doi}}

Open DOI About Zenodo DOI

Close