Moving Object Detection on CCTV Surveillance Using the Frame Difference Method

Abstract

In today's computer vision research, many build systems for observing humans and understanding their appearance, activities, and behaviour that provide sophisticated interfaces for interacting with humans, and create plausible human models for various purposes. This paper presents a simple algorithm for detecting moving objects from a static background based on frame differences. First, the first frame is captured via a static camera such as Closed Circuit Television (CCTV) after which a sequence of frames is taken periodically. Second, the absolute difference is calculated between successive frames and the difference in images is stored in the system. Third, the difference image is converted into a grey image and then translated into a binary image. Finally, morphological filtering is carried out to remove noise. In the last process, moving objects can be detected in conditions that do not change much apart from moving objects.

Penerapan Algoritma Deteksi Tepi Canny Menggunakan Python Dan Opencv

“…Kita dapat memilih kriteria seleksi untuk sampai pada hasil yang diinginkan. Kriteria yang paling umum adalah beberapa bentuk angka probabilitas dan beberapa bentuk ukuran yang tumpang tindih (misalnya Intersection over Union) [12].…”

Analisis Performa Deteksi Objek Bergerak pada Algoritma Background Subtraction dan Algoritma Frame Difference

“…Algoritma background subtraction menggunakan teknik pengurangan (selisih) antara frame sekarang (fi) dengan frame background-nya [5]. Sedangkan algoritma frame difference menggunakan teknik pengurangan (selisih) antara frame sekarang (fi) dengan frame sebelumnya (fi-1) [6].…”

“…Algoritma frame difference menggunakan teknik pengurangan (selisih) antara frame sekarang (fi) dengan frame sebelumnya (fi-1) [6]. Pada algoritma ini, selisih tidak dilakukan pada frame latar (background) melainkan pada frame sebelumnya sehingga tidak ada pendefinisian untuk frame awal.…”