Research on Intelligent Recognition System of Traffic Image Based on CNN and Intelligent Recognition of Foreign Object Intrusion

The development of road transportation is rapidly changing, and computer vision represented by image intelligent recognition has gradually become one of the mainstream research directions for the safety of transportation system operation. The article realizes the intelligent detection of foreign object intrusion by calculating the sensitive grid region of the video image, using digital image processing methods for the initial detection of moving objects in real time, and then conveying the detected abnormal video images to the CNN-based foreign object intrusion recognition model for a more accurate intelligent recognition analysis. The accuracy of the designed moving object detection method for detecting pedestrian intrusion in the experimental video is 92.33%, which is higher than the comparison method. Meanwhile, the proposed foreign object intrusion intelligent recognition method has a foreign object recognition accuracy of more than 88% in different complex environments, and overall its average accuracy is 2.04%~18.64% higher than the comparison method, presenting better recognition accuracy and recognition speed, taking into account the accuracy and real-time, and being able to satisfy the needs of the actual use of the road traffic scene.