In this master's thesis, a software solution for a system for displaying driver profile data of vehicles in front of the driver's own vehicle, based on the detection and recognition of vehicle license plates, is developed. The system consists of two applications – an application for detecting vehicle license plates and recognizing characters on them, and a web application for displaying profile data and reviewing drivers. Various implementations of solutions for detecting license plates and recognizing characters on them have been investigated, and the YOLO algorithm was selected as the most suitable for detection. The YOLOv5 small model for detection with a confidence threshold of 0.5 was selected as the most suitable for this use case, with achieved results of precision 0.9905, recall 0.9811, F1 measure 0.9858, and 2.69 processed images per second when testing on an Intel(R) Core(TM) i7-8550U CPU @ 1.80GHz 1.99 GHz processor. The PaddleOCR algorithm proved to be the most accurate for character recognition. For the maximum distance at which the final system will recognize license plates, the width of the detected bounding boxes of license plates greater than or equal to 80 pixels of the estimated maximum distance of approximately 8.13 meters at a resolution of 1920x1080 pixels was selected, where after testing PaddleOCR achieved an accuracy of 77.08%. A Django web application has been developed that enables the display of profile data and driver reviews. By testing the functionality of the web application, it was found that the behavior of the web application does not deviate from the expected behavior, except if the recognition of the license plate is incorrect, after which the page of the incorrect driver is opened. Progress in the technology used for detection and recognition or the invention of a new one opens up possibilities for improving the system presented in this paper. In addition, recognizing characters on images with higher resolution would result in higher recognition accuracy for the current system. The inability to recognize Croatian diacritics could be solved by additional training of existing models of character recognition algorithms or by developing a proprietary character recognition algorithm. The ideal dataset for testing a license plate recognition algorithm at various distances from vehicle license plates would be one that contains as many images as there are distances being tested for each license plate. The functionality of rejecting the review page that is opened based on the incorrect recognition of the license plate or the functionality that opens the review page only if the same tag is recognized with the help of a circular buffer on 5 or more images are possible improvements to the system for the problem of incorrect recognition of the license plate.