Abstract: In order to solve the problem of imprecise layering of fertilizer shovels during the fertilizer application operation of wheat root breakers, a fertilizer shovel precision row alignment system based on OpenCV was designed. With the Jetson Nano as the processor, the image information collected by the camera is processed by OpenCV, which can accurately find the center coordinate position between the wheat Long rows and thus obtain the relative position of the fertilizer shovel in each application layer. The obtained position information is passed to STM32F103C8T6 for further processing, and the position of the layered fertilizer application shovel is then adjusted through the corresponding control system. In this paper, the coordinates of the pixel points obtained by the camera are compared and analyzed with the actual coordinates measured manually, and the results show that this recognition method has a good recognition effect with a correct rate of more than 90% for the coordinates of the center of wheat under the condition of sufficient light intensity in all cases. This identification method provides a reliable theoretical basis for the design of fertilizer treatment device of root cutter in the future

Key words: fertilizer shovels, precise alignment, OpenCV, image recognition, STM32

摘要： 为解决小麦断根机施肥作业过程中施肥铲施肥不精准问题，设计一种基于OpenCV的施肥铲精准对行系统。以Jetson Nano为处理器，将摄像头采集到的图像信息通过OpenCV进行处理，能够精准找到小麦垄行之间的中心坐标位置，以此来得到每个施肥层中施肥铲的相对位置，将得到的位置信息传到STM32F103C8T6再次进行处理，通过相应的控制系统进而调整分层施肥铲的位置。将摄像头所取得的像素点坐标与人工测量的实际坐标进行对比分析，结果表明，此识别方法在光照强度充足的情况下，对小麦中心坐标的识别正确率均高达90%以上，具有较好的识别效果。此识别方法为以后断根机施肥处理装置的设计提供了可靠的理论基础。

关键词: 施肥铲, 精准对行, OpenCV, 图像识别, STM32