PEMETAAN AKTIFITAS KONSUMEN TOKO MENGGUNAKAN METODE BACKGROUND SUBTRACTION

Abstract

This study aims to heat mapping the consumer’s movement using background subtraction techniques. The mapping is built using the coordinate information obtained from the consumer location that detected from the video where the separation of the consumer object and the background is done by background subtraction technique. Tests were performed on eleven video of consumer data activity that have different activity characteristics that were created using Microsoft PowerPoint application. Simulated activities include walking straight, staying, walking back to the path that had been passed, pacing, disturbance from another object, the influence of color, the consumer walks meet and coincide with other consumers. From the test of video discovery is obtained accuracy of 96.07% for the detection process of consumer movement, where the lack of detection process occurs due to the absence of techniques used to perform the introduction of characteristics of consumer objects. The mapping process is very much in line with the number of coordinates generated in the motion detection process, but the inaccurate detection of movement in the entrance and exit areas makes the coordinates high. By filtering with Region of Interes (ROI) in the survey area, creating disturbances in the area of doors and areas with objects that produce movements other than consumers can be eliminated.