Improve Black Ice Image Detection for Road Safety

A tiny amount of Black Ice that mixtures in into the road's surface is frequently clear or virtually undetectable. In order to detect the existence of black ice on roads, a variety of sensors, algorithms, and methods are used. Black ice detection is essential, particularly for car navigational systems and safety on the roads. The purpose of the research is to propose black ice detection method based on a combination of global and local texture features, such as grey level co-occurrence matrix (GLCM) and Local Binary Patterns (LBP), to capture both overall patterns and fine details. The proposed method is divided into four phases: Image preprocessing comes first, followed by texture feature extraction techniques like GLCM and LBP. Use the relief approach to choose the best feature. Two classifiers, linear discriminant analysis (LDA) and DT, are used to distinguish between images with and without black ice. There isn't a large collection of black ice images that can be used for this research. Consequently, the black ice image data required for the training was obtained through Google image search. According to the findings of the experiments, the proposed GLCM-LBP-LDA attained an accuracy rate of 95%, surpassing the overall accuracy of the most current techniques now in use.