Detection and counting potholes using morphological method from road video

At this time the road checking is done manually. Road checking manually still has many disadvantages, among others, requires a lot of time, requires a lot of costs, dangerous, and lacks accuracy from data information. Therefore, road checks are needed automatically. In this paper an identity system was developed road distress especially potholes, based on digital video, using digital video processing. Digital video processing has advantages in detecting objects. Applications of digital video processing include are vehicle counting, pedestrian counting, etc. The steps in this paper are pre-processing for improving image quality, segmentation to distinguish between objects and background, morphology to perfect the shape of the object, labeling and tracking to give ID and know the same object in order frame, and the last step is the calculation of the number of potholes on the road. The morphology process uses morphology dilation and then morphology closing. The shape of the structure element used is a rectangle shape, while the size of the structure elements are 7 × 7, 15 × 15, and 21 × 21. In this paper, the best trial results are shown in the size 15 × 15 with an average value of accuracy 100 %, precision 100 %, and recall 100 %.