MATLAB Code for Automatic License Plate Recognition Using Erosion and Dilation Algorithms

In this project, we employ erosion and dilation algorithms to achieve automatic license plate recognition. These fundamental digital image processing techniques operate on binary images by modifying pixel values locally to perform morphological transformations. Our implementation begins with image binarization using thresholding techniques to convert the input image into a binary format. The erosion algorithm is then applied to fill gaps between license plate characters through neighborhood pixel analysis - typically implemented using MATLAB's imerode function with structural elements like disks or rectangles. Subsequently, the dilation algorithm (implemented via imdilate) expands character spacing to facilitate character segmentation. We conducted extensive experiments with various parameter configurations including different structuring element sizes and iteration counts, ultimately achieving high recognition accuracy. The code includes optimization features such as adaptive thresholding for lighting variations and morphological operation parameter tuning. Through rigorous testing and performance optimization, we have validated the algorithm's reliability and stability in practical applications, confirming the effectiveness of our license plate recognition system.