The blistering progress of the urban area has improved the traffic jams and increased the road safety hazards and therefore for traffic control (intelligent control) is extremely important to be introduced. The traditional methods of enforcement that are mostly manual and resource-consuming solution are unable to match the dynamics of the contemporary roadways. The paper gives a detailed review of the deep learning architectures that the smart traffic monitoring and enforcement of the Auto Eye is built upon AI. We examine the state-of-the-art in major technological areas, such as real-time vehicle detection and classification, and compare the performance obtained by Convolutional Neural Network (CNN) based models, such as the You Only Look Once (YOLO) series and Single Shot Detector (SSD) and the performance obtained by newly developed transformer-based networks. In addition, this review discusses the Automatic Number Plate Recognition (ANPR) by specialized Optical Character Recognition (OCR) and certain algorithms in order to identify certain violations that include non-compliance to helmet usage, speeding, and breaking zebra crossing. Such technologies, synthesized makes possible a complete automated enforcement pipeline, with the detection of violations, and the automated issuance of e-challans. The review summarizes the existing literature, points to the still existing issues of accuracy and live implementation in various environmental conditions, and provides the way forward to the building of safer and efficient urban transportation system with the help of intelligent automation.