Research On Machine Vision Based Grain Yield Sensor

This Research is originated from the project of “The multi parameter monitoring and intelligent control system of rice-wheat combine”,which is the sub-project of National Key Research Plan “Intelligent rice-wheat combine technology and equipment development ”.Precision agriculture is the agriculture which entirely integrates information technology and agricultural production,a system which periodically,locally,quantificationally implements modernized agricultural farm operation technology and management according to spatial variation.Grain yield system is the key technology of implementing precision agriculture and the foundation of implementing agricultural precision management.Aiming at the problem of low accuracy of traditional yield monitoring system,in this article a machine vision based grain yield system is designed using camera and structure light.The measurement method based on computer vision,noise-canceling algorithm and the line structure light extraction algorithm are all studied.The main content of the article is as follows:The first chapter of this thesis introduces the background and meaning of the subject.The application of the grain yield system is firstly introduced.The construction of the grain yield system is mentioned.The research status of grain yield system and grain yield sensor at home and abroad are introduced.The second chapter of the thesis is mainly about the structure design and system construction of the grain yield sensor.Firstly the construction is introduced,then the function of the modules is introduced.The selection and theory of related parts is given.The third chapter is about the grain measurement method.First the camera model and the structured light measurement model are stipulated.The calibration test and its result are mentioned.The volume model of the grain on the scrapper is introduced,therefore the theory of the grain yield sensor is stipulated.The fourth chapter studied the pre-processing method and the extraction method of the structured light.A useful algorithm of fast canceling the noise and structure light extraction are given.The fifth chapter introduces the experiment platform and the actual experiment process and results.The mechanical structure,hardware and software design is mentioned.The selection of proposed sensor and servo are mentioned.The design of the human machine interface and the information collection system are introduced.Inner experiment platform test and actual harvest test are carried out to testify the stableness and accuracy of the sensor.The sixth chapter is the summary and vision of the thesis.The main innovation points of the thesis are summarized.The next step research direction is mentioned.The proposed grain yield monitoring system is based on computer vision,which is validated on experiment platform and in real environment.The experiment result shows that,the grain yield motoring system is accurate and stable.The relative monitoring error is within 4.27%,.The actual experiment error is between 3.65%-8.12%.