Improved Yolov8 for Detection of Intruders Using a Deep Learning Techniques

• Author(s): Olatunji Oluwadare Oluwasola ; Adeniji Oluwashola David
• Paper ID: 1710342
• Page: 1191-1198
• Published Date: 03-09-2025
• Published In: Iconic Research And Engineering Journals
• Publisher: IRE Journals
• e-ISSN: 2456-8880
• Volume/Issue: Volume 9 Issue 2 August-2025

Abstract

Intrusion detection system (IDS) is an important component of computer vision and cyber security. IDS are installed and configured to protect geographical areas from unwanted intrusion and unauthorized encroachment. conventional means of intruder detection in a farmland are quite expensive, time consuming, labor- intensive, inaccurate giving false alarm and error in their detection. In this research study, various deep learning techniques were explored to identify human intrusion within a farmland. You only look Once (YOLOV8) object detection algorithm was enhanced by adjusting the neck region and optimizing the head region in order to increase its performance. The result generated was compared to the existing YOLOV8, Regional convolution neural network and the single short detector object detection model.The improved YOLOV8 model generates a precision of 93.6%, recall value of 92.7% F-1 score of 93.8% and a mean average precision of 92.9% which shows a better model performance when compared to the other Deep Learning models. This model can be used to detect unwanted and unauthorized human intruder in a real time.

Keywords

CNN, Improved YOLOV8, R-CNN, SSD, Deep learning, IDS

Citations

IRE Journals:
Olatunji Oluwadare Oluwasola , Adeniji Oluwashola David "Improved Yolov8 for Detection of Intruders Using a Deep Learning Techniques" Iconic Research And Engineering Journals Volume 9 Issue 2 2025 Page 1191-1198

IEEE:
Olatunji Oluwadare Oluwasola , Adeniji Oluwashola David "Improved Yolov8 for Detection of Intruders Using a Deep Learning Techniques" Iconic Research And Engineering Journals, 9(2)