Highway Visibility Detection Using Hough Circle Detection and Incremental Probabilistic Neural Networks

In traditional highway engineering, there are problems such as high cost and difficult maintenance in road visibility detection. Therefore, the study combines Hough circle detection algorithm with incremental probabilistic neural network to construct a visibility detection model for highway pavement. The study first uses the Hough circle detection algorithm to perform preliminary visibility detection, and then integrates the incremental probabilistic neural network with the preliminary detection to construct a detection model. These results confirm that the data processing accuracy and precision of the detection model in the image processing process are 97.03% and 93.37%, respectively. In terms of feature classification performance, its classification ability and classification time are 93.61% and 1.13 seconds, respectively. Moreover, its visibility and error percentage in visibility detection are 510.69 m and 10.06%, respectively. Regardless of the weather conditions, the environmental classification accuracy of the model remains above 90%, with the highest accuracy reaching 93.7% for sunny days. These results indicate that the highway pavement visibility detection model can improve the accuracy and stability of highway pavement visibility detection. The research aims to provide an effective visibility detection method for highway traffic safety. © 2024 Slovene Society Informatika. All rights reserved.