Research On Object Detection And Visual Navigation Technology Of Leafy Vegetable Harvesting Robot

Our country’s vegetable sown area is large and the agricultural production and operation model has changed,therefore agricultural robots have gradually become more intelligent.Machine vision technology has good applications in modern agriculture.Machine vision is an important part of robot intelligence.In terms of robots for vegetable harvesting operations,robot target detection and automatic navigation technology occupy an important position,which is important for robot intelligence.The degree is decisive.Therefore,this paper studies leaf vegetable recognition,positioning and navigation methods serving harvesting robots based on machine vision,so as to improve the level of intelligence of leaf vegetable harvesting machinery.This paper proposes a machine vision system scheme with monocular vision auxiliary line laser.According to the vision scheme,the calibration of the vision system is completed and a leafy vegetable root positioning algorithm is proposed,and the feasibility and rationality of the positioning algorithm are verified through experiments.At the same time,in order to satisfy the automatic navigation function of the robot during harvesting operations,this paper proposes an algorithm for extracting the baseline of the robot automatic navigation,and conducts the corresponding experimental research.First,according to the agronomic parameters of leafy vegetable harvesting and the performance indicators of the robot vision system,the machine vision system scheme determination and system design of monocular vision combined with line laser are completed,and the control block diagram of the overall scheme of the vision system is given.The hardware vision system was initially built,and the mathematical model of the vision system in the robot harvesting operation was established.On this basis,the principle and formula for solving the three-dimensional coordinates of any point on the laser stripe are derived.It provides a corresponding theoretical basis for the subsequent height determination of leafy vegetables.Secondly,the vision system composed of monocular vision combined with line laser was calibrated jointly with vision system parameters.Based on the calibration of the camera’s internal parameter matrix,distortion coefficient and position parameters of the laser plane,the distance measurement accuracy of the visual ranging system in different target segments was verified by experiments.Thirdly,this paper proposes an improved navigation line extraction algorithm,and conducts experimental research on the effectiveness,accuracy and real-time performance of the algorithm.A navigation line extraction technology based on the improved DBSCAN algorithm is proposed.Perform preprocessing operations on the leafy vegetable row images collected by the camera,and conduct research on target detection algorithms based on color and shape features on the basis of preprocessing operations.On the basis of effective target recognition,the algorithm of target center point extraction and navigation line fitting is researched.Contrast experimental research on the navigation line extraction accuracy and extraction speed of the improved algorithm and the original algorithm.Finally,build a hardware experimental platform to verify the image processing algorithm and positioning algorithm.Completed the experiment of vision system target detection and navigation line extraction,compared the traditional and improved navigation line extraction technology based on DBSCAN algorithm,and analyzed the target detection technology of the ranging system.In order to verify the feasibility and advancement of the proposed navigation line extraction algorithm,a comparative experimental study of the two algorithms was carried out.