Performance Analysis of Illumination Invariant Change Detection Method for Detecting Image Change in Night Vision Camera

Abstract

At present, the use of video cameras is not only limited to documenting events but is also used for surveillance systems. Changes in lighting that occur in the surveillance area is one of the problems that result in a false alarm on the surveillance system. Illumination Invariant Change Detection is a method for detecting image changes on images. This study aims to determine the performance of the Illumination Invariant Change Detection method to detect image changes in night vision surveillance cameras. The Illumination Invariant Change Detection method does not work well for detecting image changes on a night vision camera under dark lighting conditions at an average value of Lux 0 with an infrared lamp on. The accuracy of the application of the method to detect image changes on night vision cameras is 80% with the selection of the threshold value of the detection of image changes that is 75000 pixels.

Perancangan Sistem Keamanan Pintar Kamera Night Vision Auto Color berbasis Raspberry PI

“…Cahaya yang berasal dari bulan, bintang, atau lampu yang ada di sekitar objek merupakan sumber yang bisa ditangkap kamera night vision. Perubahan pencahayaan menjadi salah satu kendala dalam berbagai deteksi dan identifikasi pada sistem pengawasan yang memanfaatkan kamera video [10]. Pencahayaan yang ekstrim tentunya akan menimbulkan sistem pengawasan menjadi tidak akurat dari perubahan sistem mendeteksi adanya objek pada kenyataannya objek tersebut tidak ada.…”