Road lane markings are one of the main factors that cause traffic accidents and jeopardize the safety of both drivers and pedestrians. Both computer vision systems find it challenging to recognize road lanes. The irregular maintenance, insufficient drainage, and traffic patterns. It negatively affects driving comfort, road safety, and vehicle condition, and it could also lead to a lot of accidents. Recently, several machine learning algorithms have been used, although their accuracy and efficiency have not improved significantly. We suggest applying Region-Based Convolutional Neural Networks (RCNN) for path identification to tackle this issue. RCNNs are another kind of neural network that is capable of extracting significant information from picture and time series data. This makes it highly helpful for tasks requiring images, such as shape identification, object classification, and recognition. Image from a dataset with noise removed using a Wiener filter. The next step is to extract features from images using Linear Discriminant Analysis (LDA). Determine which lane line in the RCNN is efficient and provides accuracy, precision, recall, and F1-score as 95.6%, 94.05%, 94.58%, and 94.23%, respectively.
Road Lane Detection using Region-based Convolutional Neural Network (RCNN)
2025-01-24
345821 byte
Conference paper
Electronic Resource
English
Convolutional neural network automobile lane changing auxiliary system
European Patent Office | 2025
|