<b>A HAAR CASCADE AND FRAME DIFFERENCING APPROACH FOR SELECTIVE PLANT MOTION DETECTION IN SURVEILLANCE SYSTEMS</b>

This research project addresses the growing need for specialized motion detection systems capable of distinguishing vegetation movement from other dynamic objects in surveillance environments. While existing computer vision technologies excel at detecting human and vehicular motion, a significant gap remains in accurately identifying and analyzing plant and tree motion, which often triggers false positives in conventional CCTV systems. The study aims to develop an algorithm using Python and OpenCV to detect vegetation motion selectively, optimizing storage efficiency by ignoring irrelevant movements.The objectives include precise plant detection using a Haar Cascade classifier, motion isolation through frame differencing, and integration into surveillance systems to reduce false recordings. The methodology involves capturing sequential frames, computing absolute differences, converting images to grayscale and binary formats, and applying morphological filtering to minimize noise. Despite limitations such as training on a single plant species and susceptibility to environmental interference, the algorithm demonstrates promising results in detecting partial plant movements under controlled conditions.The findings highlight the algorithm’s potential to enhance surveillance systems by reducing storage overhead and improving motion detection accuracy. Future work should expand training datasets and incorporate stabilization techniques to address environmental challenges. This research contributes to the niche field of vegetation motion analysis, offering practical implications for automated surveillance and ecological monitoring.