Wheat rows detection at the early growth stage based on Hough transform and vanishing point

• A simple and effective wheat row detection method is presented.
• A moving window and multiple interlaced scanning strategies were constructed to extract feature points.
• The Hough transform is used to estimate all possible candidate wheat rows.
• K-means clustering was performed to look for a clustering center representing the vanishing point.
• The real wheat rows were detected based on the vanishing point.

A simple and effective wheat rows detection method is presented in this paper. It includes five steps: image segmentation, feature points extraction, candidate wheat rows estimation, vanishing point detection and real wheat rows detection. Firstly, a color image was converted into gray-level image and segmented in two parts, the foreground and background. Secondly, to extract feature points indicating centers of wheat rows, a moving window and multiple interlaced scanning strategies were constructed. Thirdly, the Hough transform method was employed to extract straight lines for estimating all possible candidate wheat rows. Fourthly, k-means clustering was performed to look for a clustering center representing the vanishing point. And lastly the real wheat rows were extracted based on the vanishing point. Test results indicate that the proposed method can effectively detect wheat rows at the early growth stage, and the detected rate is up to 90%.