The Research On Coprinus Chinensis Picking Robot Binocular Vision System Identification And Location

The research of machine vision on the picking robot can help reduce the labor intensity of farmers,alleviate the shortage of labor in China,and can greatly improve the picking efficiency.The picking robot can locate the target in a short time through the visual system during the picking process.The object has a great influence on improving the working efficiency of the picking robot.There are many kinds of signal collecting devices used in the picking robot vision system,such as a camera and an infrared sensor.Considering that the picking period of Coprinus comatus is short,it must be completed within a certain period of time.The binocular vision system in this paper uses the camera as the signal acquisition device.Therefore,machine vision is one of the important contents of the research on the picking robot of the Coprinus comatus,which directly affects the chicken legs.Picking efficiency of mushroom picking robots.The development of machine vision is affected by the development of image processing methods and signal acquisition devices.Many picking robots now rely on cameras to capture images for fruit recognition and localization.However,the visual system has low fruit recognition accuracy and inaccurate positioning.insufficient.Therefore,it is of great practicality and good application prospect to study the identification and localization of Coprinus comatus through binocular vision.The work of this paper is mainly from the following aspects:1.Establish a corresponding mathematical model based on the camera-based imaging model.According to the pinhole model of the camera,the internal parameter matrix of the camera can be solved,and the distortion parameter of the camera can be obtained by the distortion model.The CCD camera was finally selected through the selection of the camera and the establishment of the calibration experimental platform,taking into account factors such as price and introduction of noise.Using the obtained camera mathematical model,the calibration experiment is finally carried out;the camera calibration is performed with 20 calibration images using the OpenCV library,and the inner parameter matrix,the rotation matrix and the translation matrix of the left and right cameras are obtained.2.Correct the image of the left and right cameras by using the camera's parameter matrix,and use the corrected image for post-processing;improve the mean filtering by adding the gradient influence factor on the basis of the mean filtering algorithm,so as to preserve as much as possible while filtering The edge of the image and the improved image are processed by the improved mean filtering method to reduce image noise and preserve edges,which facilitates subsequent image segmentation.3.Using Canny edge detection for the corrected and noise-reduced images,then the outer contour of the koji mushroom and the edge detection of the cap are performed on the image.On the basis of the edge detection,the shape and contour of the koji mushroom are searched,and the ellipse fitting is performed.The feature point matching area is determined,and the feature points on the chicken leg mushroom are detected in the found area to perform stereo matching.In the actual picking process,there are many Coprinus comatus in each image,and the corresponding Coprinus comatus can be determined according to the feature point matching.Use the left and right camera images to obtain the minimum circumscribed rectangle and elliptical contour of the genus Coprinus comatus cover to determine the actual size and position of the cap,and judge whether the Coprinus comatus is mature according to the size of the cap,and then decide whether to pick.