Detasseling of maize at tasseling stage is an important task to increase maize yield,but the color of maize leaves,tassel and soil is difficult to separate,which makes it difficult for the Maize Detasseling Machine to navigate between crop rows.Considering that the key to navigation line extraction is the detection of crop feature points in the driving area of agricultural machinery,this paper independently extracts the ROI of maize male ear through YOLOv8 neural network,and uses Jet mapping and Otsu algorithm to segment maize tassel,green leaf and soil within the ROI.Then,the FAST corner detection method was used to extract the male tassel feature points and divide the feature points set according to the navigation region.Finally,the least square method was used to simulate the detection line of the cooperative object.The experimental results show that the proposed algorithm can achieve accurate and fast navigation line extraction during maize tasseling stage.The average time of processing a single frame image(600 pix × 700 pix)is 56.7 ms,and the average deviation Angle is 1.03°,which can meet the requirements of Maize Detasseling Machine navigation in the field.
维普
万方数据
