| In China,agriculture has always been regarded as the national pillar industry,the development of modern agriculture automation is a major historic task in the new era of China's modern agriculture.Modern agriculture in China can only survive in the direction of highly mechanized and automated development.In recent years,with the development of information technology,artificial intelligence technology in agricultural machinery has been applied more widely.Agricultural machinery equipment presents the trend of information technology intelligence.In order to improve the level of China's cotton mining automation,this paper designs and develops the intelligent perception and analysis platform for the integrated multi-sensor cotton picking machine by using the technology of sensor,computer vision,GPS and GIS in the intelligent research of agricultural machinery,to form an intelligent equipment to effectively enhance the automation level of the cotton picker.The main research contents and achievements of this paper are as follows:(1)Research on entire information collection method of cotton picker.This paper has analyzed the information needed to realize the entire information monitoring of the cotton picking machine,and broken through the limitation of the traditional agricultural machinery monitoring information limited to the state of the agricultural machinery through the sensor,camera and external GPS and other sensors integrated approach.It has realized including the information of the cotton picker,harvesting environment,time and space information of the whole information collection and storage,for the realization of truly intelligent monitoring equipment to lay the data foundation.(2)Research on Key Technologies of cotton picker automation.Solve some key technology of agricultural machinery automation includes visual navigation along the line,cotton warehouse full warning,calculation of cotton yield.Visual navigation along the line.Taking the front view image of the cotton picker machine as the analysis object,the computer vision and GIS spatial analysis are the main technical means to realize the automatic calibration of the cotton orbital track based on the target recognition and the spatial analysis,cotton boundary extraction based on cotton seed point extraction,morphological expansion and gap filling,automatic fitting of boundary polygon in cotton field based on adaptive distance parameter dynamic programming,and the automatic alarm of the picking deviation of the picking machine has been realized by the method of GIS spatial analysis.Cotton warehouse full warning.Taking the image of cotton bones as the object of analysis,the idea of intrusion detection has been brought into the warning of the cotton warehouse.By drawing the cotton support column warning line and the binary image of cotton bones based on histogram equalization,bilateral filtering and Otsu threshold segmentation,a decision-level cotton warehouse full warning has been realized.Calculation of cotton yield.Taking the trajectory points of picking machine as the data base,the buffering topology fusion technology in GIS as the main technical means to achieve a high precision calculation of the number of cotton picking through the trajectory points to restore the spatial distribution of cotton fields.The accuracy and reliability of the experiment have been verified by simulation experiments.(3)Research and development of intelligent monitoring platform.This paper has designed and implemented an independent,intelligent cotton picking machine intelligent perception platform.The platform has realized the entire information collection and display of cotton picker condition,harvesting environment,time and space information.The platform has implemented many intelligent analysis functions including visual navigation,full warehouse detection,cotton acreage calculation,to ensure the process of picking cotton,safe,reliable,efficient,and to effectively improve the level of China's independent equipment cotton picker machine. |
PDF Full Text Request