Research On The Technology Of Cotton Recognition And Localization Based On Machine Vision

Xinjiang is the biggest quality cotton production base in China, cotton picking is the key to harvest the cotton harvest in mid October every year in.. In order to improve production efficiency and reduce production costs, the use of mechanized cotton wool. But because of the way for a one-time job and need cotton maturity to the same or similar, so need in the picking operation before the ripener spraying. At the same time, in order to prevent the cotton impurity in the process of picking, the quality grade of cotton, need to remove the leaves from the leaves before picking the job.. This use of pesticides not only increased the cost, and caused great harm to the soil. Therefore, it is a key problem to study the automatic identification and localization of cotton in the cotton picking machine which is mechanization and intelligence.. In this paper, using machine vision technology for cotton characteristics, design and build cotton identification and positioning system, the work of this paper is as follows:1, through to the domestic and foreign mining status of cotton and machine vision technology on the progress of the research. Through the analysis of the content, target and technical route.2, the application of mechanical design theory and the theory of machine vision, the design of cotton recognition and localization system based on machine vision and more actual situation choose suitable experimental platform, and the system is debugged.3, the analysis of CCD camera calibration principle, using the calibration board, and calibration toolbox to achieve the calibration of CCD camera, required to achieve the vision sensor CCD camera internal and external parameters.4, according to the research target, the cotton picking by digital camera and a CCD camera acquisition mature cotton static and dynamic image, and using the technology of image processing to complete the cotton the image preprocessing.5, maturity of cotton images in different color space segmentation segmentation mature cotton image, extraction of mature cotton image features, the PNCC coupled neural network algorithm in image segmentation.6, by using centroid picking point matching method, realization of mature cotton left image and the right image centroid picking point for stereo matching, at the same time, into the front of the calibration of camera parameters, calculated to pick cotton picking the 3D coordinates of point space.7. Based on the research of spatial localization of cotton, the technology route of the whole scheme of the fruit recognition and locating software system for the mature cotton is designed.. In the integrated development environment of C++ Visual, the feature extraction system of mature cotton is developed with r2012 b GUI MATLAB toolbox, and the running and debugging of cotton recognition and positioning system is carried out..This topic according to cotton picking period cotton picking machine imitate human picking when visual system function, so as to solve problem of identification and location of cotton, built the cotton recognition and localization system based on machine vision technology, the recognition and location of mature cotton. The topic for the research of intelligent of cotton picker laid the foundation.