DETECTION OF ILLEGAL ON-STREET PARKING

Abstract

The increase of automobiles in this generation is very high compared to the olden days. Due to the increase of such automobile, there is also a gradual increase in terms of traffic in the countries. We wait for a long time based on the traffic in the cities. The main cause for some traffic to occur is due illegal parking on roads and other areas. In order to beat this drawback, we tend to propose a way to avoid traffic that is by detecting illegal on-street parked vehicles by the use of their number plates by image processing. On-street parking refers to practice of parking vehicle on the side of public road or street. The on-street parking is common method of parking in urban and suburban areas where parking lots or dedicated off-street parking space may be limited. Due to illegal on-street parking there will be traffic congestion and also accidents may occur. “Detection of Illegal On-Street Parking” provides data of the illegally parked vehicle by capturing the image of number plate and that image is converted to greyscale, segmented and masked in the pre-processing stage. Then using Computer Vision technique, bounding boxes are formed around characters and characters are also printed with the use of CV library. The proposed also generates the chalan or invoice of penalty stating the duration of time parked in. Key Words: on-street parking, area of interest, Raspberry Pi, image processing, Resnet, Yolo v3 model, OCR.