PERBANDINGAN DETEKSI OBJEK PADA IMAGE SEQUENCE MENGGUNAKAN HAAR CASCADE CLASSIFIERDAN BACKGROUND SUBTRACTION
Theft is a classic problem from ancient times to the present that can occur in various places such as parking lots, houses, farms to agricultural land. Various efforts to prevent theft have been carried out by both the community and the police, such as increasing patrols, siskamling and even other efforts from the community by installing Closed Circuit Television (CCTV) surveillance cameras, but these efforts are sometimes felt not optimal. Therefore, in this study the author tries to apply a background subtraction method for detecting moving objects in the video. The stages of the process of applying the background subtraction method in detecting moving objects in the video are finding the height, width and video frame, creating a matrix to accommodate the frame, extracting each RGB value, finding the mode value for each RGB value, combining the mode values for each RGB value (finding the background). , converts the RGB image to grayscale, reduces the value of the frame with the background (object detection). In this study, the detection of moving objects using the background subtraction method was carried out in two conditions, such as morning and evening. From the results of testing the background subtraction method for all conditions, the accuracy in detecting moving objects reaches 95%. As for testing the haar cascade classifier, it is still in the testing phase.