Short Communication
Year: 2021 | Month: September | Volume: 8 | Issue: 9 | Pages: 130-135
DOI: https://doi.org/10.52403/ijrr.20210918
A Remote Surveillance System for the Detection of Farmland Invasion by Cattle in Sub-Saharan African Country
Oluwole Abiodun Adegbola1, Ifeoluwa David Solomon2*, Adesina Samuel Oluwaseun3, Odebunmi Nathaniel Dare4, Peter Olalekan Idowu5*
1Lecturer I, Department of Electronic and Electrical Engineering (EEE), Adoke Akintola University of Technology (LAUTECH), Ogbomoso, Oyo State, Nigeria.
2*Research Assistant Signal and Information Processing Instrumentation and Control (SIPIC) Research Group, Department of EEE, LAUTECH, Ogbomoso, Oyo State, Nigeria.
3Former Student Department of EEE, LAUTECH, Ogbomoso, Oyo State, Nigeria.
4Former Student Department of EEE, LAUTECH, Ogbomoso, Oyo State, Nigeria.
5*Research Assistant SIPIC Research Group, Department of EEE, LAUTECH, Ogbomoso, Oyo State, Nigeria.
Corresponding Author: Peter Olalekan Idowu
ABSTRACT
Farmer-herder conflict in Nigeria mainly involves disputes over land between agrarian communities and nomadic Fulani herdsmen resulting to loss of farmlands and crops, which consequently affects the nation's economy. Engagement of local security operatives in stopping this menace of the herders have proved abortive. Hence, this work proposes a farmland surveillance-alert system using unmanned aerial vehicles for the detection of cattle presence on farmlands as a solution to curbing the problem of farm invasion and destruction. The technique modifies CNN-YOLOV2 architecture, the outcome was accessed with DJI phantom 4 captured 656 images for the detection of cattle invasion. The system on detecting cattle presence above a threshold level sends SMS to farmer’s designated number. The system achieved an average confidence score of 0.92 for the test dataset and 0.72 on real-life data. Hence, it can be employed to mitigate incessant farm invasion and destruction problem and in other surveillance systems.
Keywords: convolutional neural network, unmanned aerial vehicles, surveillance, mean average confidence score, yolov2, short messaging service.
[PDF Full Text]