Monitoring Method And Device For Impurity Rate And Crushing Rate Of Flowing Grain Based On Multi Threshold Segmentation

With the improvement of agricultural mechanization level in China,the combine harvester has been widely used.Most of the field crop harvest is completed by machine instead of manpower,which greatly improves the efficiency of harvest.The impurity and crushing rate of the harvested grain are important reference indexes to measure the quality of the grain and the performance of the combine harvester.The real-time monitoring of the grain can help the harvester driver understand the operation situation,adjust the operation parameters according to the situation,raise the efficiency of the operation and the quality of the harvested grain.At present,the foreign combine harvester has deployed the real-time monitoring system of impurity and crushing rate,which can monitor the grain impurity content crushing rate in real time when the harvester is working;But there are many sampling steps and some problems in realtime monitoring for the monitoring of the impurity and crushing rate of the combine harvester in China.In view of the above problems,this paper studies the monitoring method and device of the broken rate of the mobile grain based on multi threshold segmentation.The real-time monitoring of the impurity and crushing rate is realized by collecting the moving grain image of the grain auger flowing into the grain tank by the combined harvester and processing the image with multi threshold segmentation algorithm.The images and processing results of the device can be displayed on the display screen in the cab.The driver can understand the impurity content and crushing rate of the grain harvested through the display interface and adjust the operation parameters according to the basis.At the same time,the device can transmit data to can bus of combine harvester for other airborne equipment;In addition,the acquired image and the corresponding processing data are stored locally for subsequent research.The main contents and conclusions of this paper are as follows:1.Establish the calculation model between the pixel area of grain image,the impurity content and the crushing rate.The process of detecting the impurity and crushing rate by calculating the mass ratio is analyzed,and the relationship between image pixel and quality is calibrated.According to the method of calculating the impurity and crushing rate manually,the calculation model between the pixel area of image and the ratio of clutter and fragmentation is established.2.Hardware selection and construction of monitoring device for impurity and crushing rate of mobile grain.The paper analyzes the movement process of grain in the operation of the combine harvester,and puts forward a monitoring scheme to guide the grain logistics movement and collect image processing at the outlet of the grain conveying auger of the combine harvester.The monitoring device consists of the image acquisition device installed under the grain conveying auger in the grain box and the main control device placed in the cab.The image acquisition device includes grain sampling module and image acquisition module.The grain sampling module completes the random sampling of grain by guiding the grain at the grain outlet to flow through the device.The image acquisition module collects the grain image flowing in the device.After the development of the device,the feasibility of the image acquisition is verified on the test bench.According to the monitoring requirements,the appropriate main control unit,power module and can signal converter are selected and packaged in the main control device box to realize image processing,signal transmission,data storage and display functions.3.Research on rice image processing algorithm based on multi threshold segmentation.Firstly,the image is enhanced,and then the RGB distribution difference and shape area difference of the whole rice grain,broken rice grain and impurity are studied.The algorithm of grain image processing is designed by setting R,B and area threshold by difference setting.The designed algorithm is used to identify 200 images collected by the image acquisition device,and the results are quantified.F1 value is used as the quantitative index of pixel level evaluation of recognition results.The F1 values of broken grain,rice stalk impurity and rice stalk impurity are 92.92%,90.65%and 90.52%,respectively,The image recognition algorithm can effectively identify the broken grains and impurities in the image.4.The software design of monitoring device for the impurity and crushing rate of mobile grain based on multi threshold segmentation.The software functions include the functions of controlling the industrial camera to complete grain image acquisition,grain image processing,data local storage,can data transmission and interactive interface display.The image acquisition program drives the industrial camera to obtain the image by exposure according to the preset exposure parameters.The image processing program uses the multi threshold image processing algorithm and the calculation model of the rate of clutter to process the acquired image,and calculates the clutter and the crushing rate of the image.The local data storage program saves the impurity and fragmentation rate of real-time processing to local for subsequent analysis and research;Can data transmission program sends data to can bus of combine harvester and communicates with other hosts.5.The experimental research and design of the device for the detection of the rate of fragmentation of the impurity content based on multi threshold segmentation.The performance of the on-line monitoring device for the crushing rate of impurity content was tested through indoor and field tests.The relative errors of the impurity content and crushing rate were 9.52% and 9.26% respectively,and the average single monitoring time was about 2S.The field experiment deployed the monitoring device on the combine harvester,and the crushing rate of rice was monitored in real time during the operation of the combine harvester.The monitoring period was 2S,the relative error of the impurity content and crushing rate was 9.72% and 9.39%.The precision and realtime performance of the device met the expected effect.