build me an automatic vehicle license plate recognition software robust to partial occlusion

the work involves development of a Nigerian license plate recognition that tackles the issue of partial occlusion and other background complexities. INSTRUCTIONS 1. The software must be done on a Windows OS. 2. C++, Python programming language is preferred. 3. The developer might use any algorithm/technique of his choice for the Pre-processing Phase, Localization Phase and Segmentation Phase but the use of Convolution Neural Network is Preferred for this task. 4. Edit the images to form an artificial covering for partial occlusion recignition 5. The software must show: * Dialog to load Vehicle images * Each phases of the image recognition, * Recognized character with % of other predicted license characters, * Recognition time, * Evaluations (i.e error rate and accuracy). * A minor Report explaining the algorithm used in each phases. Data set of about 500 vehicles are available for Testing and training. Thanks