Abstract:

The quantification process of the roadside landscape can be divided into 2 stages. The first stage is the image segmentation. Three segmentation methods such as marker-based watershed algorithm, Canny operator edge detection algorithm and texture analysis algorithm were comparatively analyzed, and the texture analysis algorithm which is characterized by more robust was selected. The second stage is the feature extraction of color space and continuity. In the extraction of the landscape color space, using the k-mean cluster analysis to extract the RGB vector of landscape, the color segmentation system was established making reference to the Munsell's color segmentation method, and the macro-color of the landscape was identified with high accuracy. In the extraction of landscape continuity, taking the landscape of section of Kunyuan freeway as a case example, the HSV values based on time series were extracted. The relationship between the HSV color space and the roadside landscape was analyzed quantitatively and the results show strong correlativity.

Key words: engineering of communications and transportation safety, roadside landscape, texture analysis, image segmentation, feature quantification, k-mean cluster