License Plate Detection and Integral Intensity Projection for Automatic Finding the Vacant of Car Parking Space

Abstract

This paper demonstrates approaches for automatic counting the number of car in a parking lot image by using the location of license plate area and the result of intensity projection. Firstly, in order to obtain the car height boundaries, we applied the edge processing histogram in a vertical direction. This histogram results from the total summation of difference intensity values between the pixel and it neighboring pixel of the enhanced image. Next, the result of derivative method is applied to segment the image into the region of interest (ROI). To estimate the license plate area, the edge processing histogram is utilized again along the horizontal direction. After that the integral intensity projection is applied to sum and smooth both intensity projection result of the histogram and the ROI image in vertical direction. Finally, the number of parking spots were detected and counted by using peak detection results. The accuracy for counting the number of cars in the indoor parking lot image is 81.16% by applying these techniques. © 2019 IEEE.