Research On Recognition And Orientation Method Of Sugarcane Buds Based On Machine Vision

With the development of intelligent agricultural machinery,visual inspection technology has been widely used in the field of agricultural machinery,and crop identification and orientation methods have become the research hotspots.Sugarcane is the main sugar crop in our country.It has now become the largest strategic economic crop in Guangxi and the main source of income for farmers.When sugarcane is planted,the sugarcane buds need to be directed towards the agronomic requirements of the ditch wall.At present,the directional adjustment of sugarcane seeds is completely dependent on human hands to plant sugarcane,and the technology of automatic sugarcane planting has become a bottleneck for the rapid development of intelligent agricultural machinery.In this study,sugarcane seeds were used as the test object,and a recognition and pose adjustment system of sugarcane buds based on machine vision was designed,this paper studies recognition and pose adjustment algorithms of sugarcane buds.The concrete research content is as follows:(1)This paper designs and builds an image acquisition system of sugarcane seed for visual inspection,analyzes and determines the types and models of industrial cameras,lenses and light sources.In order to eliminate the adverse effects of lens distortion on the image collection of sugarcane seeds,the “Zhang’s Calibration Method”is used to calibrate the camera.(2)The median filter and Otsu threshold segmentation method are used to segment and extract the multiple types of region of interest in sugarcane seeds,which can effectively filter out region of background noise and non-sugarcane bud region.A labeling algorithm based on connected domains is proposed,which extracts 5characteristic parameters from the color and shape features of the region of interest,and participates in the subsequent research on the pose recognition of sugarcane seeds.(3)In order to realize the accurate recognition of sugarcane buds,a recognition algorithm for sugarcane buds based on fuzzy theory and wavelet transform technology is proposed.Through the analysis of the image of the sugarcane seeds under the H component,the image of the sugarcane seeds are processed using morphological operations such as corrosion and negation.Based on the fuzzy theory and relief algorithm,the image of the sugarcane seeds under the H component are enhanced,and the sugarcane buds area is further highlighted,and the contour of the sugarcane buds are extracted based on the edge detection algorithm of wavelet transform.Using image multiplication and morphological operations to remove the noise of the sugarcane buds outline and the sugarcane buds area,the results of the detection test for the surface with and without sugarcane buds are: the accuracy of the recognition of the surface with sugarcane buds is 92.5%,and the surface without sugarcane buds is 89.2%.(4)In order to recognize the sugarcane buds position and sugarcane seed orientation,a sugarcane seed position recognition and orientation adjustment method based on the sugarcane buds position is proposed.The support vector machine classifier is used to classify the non-edge and edge regions of interest respectively,and the classification accuracy is 94.8% and 91.7% respectively.The position of the sugarcane buds was used to analyze the position of the sugarcane seeds,and the space model of the sugarcane seed’s position is established,and the deflection angle of the sugarcane seed’s position is calculated.An optimized pose recognition algorithm of the sugarcane seeds based on decision tree is proposed to avoid misrecognition of non-edge and edge positions using two different support vector machines.(5)After completing the position recognition of the sugarcane seeds,the orientation adjustment angle of the sugarcane seeds is calculated according to the information of the deflection angles of the sugarcane seeds in different positions and states,then the position adjustment plan of the sugarcane seeds is determined,and finally the movement of the orientation adjustment of the sugarcane seeds in different positions is completed.The test results show that the success rate of the directional adjustment of the sugarcane seeds is 94.5%,which can meet the agronomic requirements of the directional planting of sugarcane seeds.