An automatic detection method to the field wheat based on image processing

摘要

The automatic observation of the field crop attracts more and more attention recently. The use of image processing technology instead of the existing manual observation method can observe timely and manage consistently. It is the basis that extracting the wheat from the field wheat images. In order to improve accuracy of the wheat segmentation, a novel two-stage wheat image segmentation method is proposed. Training stage adjusts several key thresholds which will be used in segmentation stage to achieve the best segmentation results, and counts these thresholds. Segmentation stage compares the different values of color index to determine which class of each pixel is. To verify the superiority of the proposed algorithm, we compared our method with other crop segmentation methods. Experiment results shows that the proposed method has the best performance.