Research On Machine Vision Technology Of Caragana Stubble Machine

As an important ecological crop in Ningxia,Caragana korshinskii not only plays a huge role in preventing wind and sand fixation,but also has excellent feeding value.However,due to its poor growth environment and complex harvesting operations,its current flat stubble method is single,the flat stubble effect is poor and the harvesting efficiency is low,and its ecological and environmental benefits and economic use value are not fully utilized.In order to improve the stubble effect,this paper uses machine vision technology to solve the problem of adaptive terrain imitation of the stubble header,and realizes the refined and intelligent operation of the stubble stubble,which has important practical engineering research value.In this paper,based on the development status of korshinski stubble machine at home and abroad and the application of machine vision technology in agricultural engineering,a machine vision system scheme with monocular vision assisted line laser is proposed.According to the vision scheme,the calibration of the vision system was completed,and a positioning algorithm for the roots of Caragana korshinskii based on distance measurement information was proposed,and the feasibility and rationality of the positioning algorithm was verified through experiments.Based on the mechanization of harvesting equipment,To realize the intelligent stubble operation of Caragana.First of all,according to the agronomic parameters of Caragana korshinskii stubble and the performance indicators of the stubble machine,the machine vision scheme and system design with monocular vision combined with line lasers were completed,the overall scheme flow chart of the vision system was given,and the hardware vision was initially built The system establishes a mathematical model of the vision system during flat stubble operation,and uses the principle of triangulation to derive the three-dimensional space coordinates of any point on the laser stripe,which provides a corresponding theoretical basis for the subsequent positioning of the root of Caragana korshinskii.Secondly,the camera and vision system were calibrated separately for the vision system composed of monocular vision combined with line laser.In the process of using the calibration toolbox in the numerical analysis software to calibrate the camera,the internal and external parameters of the camera and their corresponding distortion coefficients are determined.Using a high-precision spiral displacement platform with an accuracy of 0.02 mm,the depth information and height information of the vision system are calibrated,the deviation coefficient of the vision system is solved,and the derived three-dimensional space coordinate formula of the root is further optimized,In order to improve the positioning accuracy of the vision system..Thirdly,image recognition and positioning of the root branches of Caragana korshinskii.According to the color characteristics of the collected images of the branches and trunks of Caragana korshinskii,it is preprocessed by brightness unification,gray-scale,image filtering,and then segmentation processing to obtain a binary image with only laser stripes,which is finally extracted by the skeleton thinning algorithm Single pixel laser stripes.Combining the principle of triangulation distance measurement and the results of system calibration,a positioning algorithm of Caragana korshinski root coordinates based on distance measurement information is proposed.The roots of Caragana korshinskii are located,and the three-dimensional coordinates of the roots of Caragana korshinskii are obtained.Verify the correctness of the algorithm.Finally,build a hardware experimental platform to verify the image processing algorithm and positioning algorithm.By comparing the actual measurement results with the theoretical calculation values,the corresponding error sources are analyzed.Experiments verify the feasibility of the proposed dynamic image processing algorithm based on inflection point features,and provide technical support for the intelligent control of the adaptive header.