Design of Enhanced License Plate Information Recognition Algorithm Based on Environment Perception

Aiming at the problem of difficulty in completely and accurately identifying number plate information in bad weather, this study proposes an algorithm based on environmental perception and the enhanced recognition of number plate information. Based on the YOLOv11 framework, we developed a histogram YOLO (H-YOLO) detector and designed an Enhanced Transformer Block (ETB) to improve the extraction of regional features of the number plates in complex environments. Meanwhile, a Real-Time Weather-Aware Image Enhancement Module is designed, which integrates real-time environment awareness technology to dynamically adjust the image enhancement strategy according to the changing weather conditions and improve the image quality. Optimization of the detection architecture was achieved through thresholding and independent image enhancement of the plate region. Comparing the H-YOLO detector with the baseline YOLOv11 model, we found that the training precision and recall of our model were improved by 2.62% and 1.8%, respectively. Furthermore, in real-world detection experiments, the H-YOLO detector was improved by 12 percentage points. The effectiveness of our proposed perceptual enhancement algorithm is further confirmed by the fact that when validating the China City Parking Dataset 2019 (CCPD 2019), it improves the detection rate by 24.53% compared to traditional image enhancement methods. In terms of computational efficiency, the new detection architecture reduced the load by approximately 9% and improved the inference speed by 15%. This study effectively improves the success rate of number plate recognition under real-time changing and complex weather conditions, and provides reliable technical support for practical applications.