Abstract:

The research on forest log transport vehicle identification can effectively prevent the abnormal behavior of illegal transportation and improve the effectiveness of forest resource monitoring and management.However,the complex road scene in the forest area and the color of the log end face are susceptible to light,humidity,etc.,which increases the difficulty of the log vehicle identification.This paper proposes a forest vehicle identification method based on YCbCr color space and Hough transform circle detection.Considering a bundle of logs has the same background and small differences in color,using brightness and color features to segment the image into a YCbCr space with an excellent real-time performance to remove background interference.The image is reconstructed to the RGB space to obtain the binary image of the log area with the background removed.The morphological method is used to uniformly remove log gaps and filter the interference pixels to determine the edge of the image accurately.The duality between the dots and lines of the Hough transform circle is used to detect the log transport vehicle,and to reduce the sensitivity of noise.The experimental results show that the recognition accuracy rate for the bundle of bare log vehicles reaches more than 71%,and the identification method has good robustness and practicability.

Key words: forest vehicle, feature extraction, YCbCr, Hough transform circle, vehicle indentification