Abstract: In order to quickly and accurately detect the centerline of the parent line of the rice production field, combined with the crop growth and the agronomic needs of the operation, a method for rapid extraction of the navigation line of the crop row during the rice pollination period is proposed. Firstly, according to the difference of RGB color components of the parents of the crop row, the RB color difference method is proposed for gray image processing, and the gray histogram is used to select the segmentation threshold to binarize the gray image; then traverse the image to mark the region of interest, use the maximum connected area method to extract the target line, and finally scan the image to extract the feature points and use small square method to fit the navigation baseline. The test results show that this algorithm can quickly and accurately extract the navigation line compared with other algorithms. The accuracy rate is more than 92%, and extracting a single picture takes about 200 ms. The reliability of the algorithm is tested in different environments. It has good robustness. This paper provides a reliable navigation method for the UAV-assisted operation of rice-pollinated pollination groups.