Motorcycles Counting Using Binary Segmented Area Estimation

This paper address the problem of motorcycles parking on the street such as irregular parking and high motorbike density. This is a challenge for researchers to find optimal solutions in conducting object detection research. This paper aims to calculate the number of motorcycles based on the estimated foreground area. Estimating the number of motorbikes at each predetermined time is very important to know the number of parking lots that are still available. The problem that arises in image processing is the limitations of the hardware system in processing and available memory, this is an obstacle in program development. One solution to optimize hardware systems is that a simple approach is needed. In this paper, the first step is done by extracting the foreground to estimate the area of one motorcycles. The pilot program was carried out in a motorcycle parking area that had a high density, the results of the trial showed that the approach has good accuracy for estimating the number of parked motorcycles with promising tolerance.