Development And Experimental Study Of On-line Monitoring Device For Grain Impurity Rate And Broken Rate Of Combine Harvester

The emergence of the integrated harvester for grain harvesting has improved the efficiency of agricultural production.The impurity rate and broken rate of grain are important parameters for evaluating harvesting quality.Some foreign intelligent combine harvesters have been equipped with on-line monitoring devices for impurity rate and broken rate,which can work steadily for a long time.Domestic combine harvesters have not yet been realized,so people can only detect impurity rate and broken rate manually.In this thesis,an on-line monitoring device for impurity rate and broken rate was developed.The device used machine vision technology instead of manual detection method,which could monitor the content of impurities and broken grains in harvested grains online and display on the display interface of the cab.It was convenient for drivers to intuitively understand the monitoring situation of grain impurity rate and broken rate.Once the data exceed the standard,drivers could check and adjust the operating parameters in time.The monitoring device transmitted data to upper computer through serial port for aggregation,which providing conditions for intelligent control of combine harvester.At the same time,the image and monitoring value were saved in SD card,which was conducive to off-line research and analysis.The main contents and conclusions of this thesis include:1.The detection methods of impurity rate and broken rate of rice harvested by combine harvester were studied,and the working principle of on-line monitoring device for impurity rate and broken rate was analyzed.According to the traditional detection methods of impurity rate and broken rate,the calculation methods of image-based impurity rate and broken rate were studied.According to the relationship between weight and area of pixels of complete rice,broken rice and impurities,the calculation models of impurity rate and broken rate were obtained by regression analysis.2.Hardware design of on-line monitoring device for impurity rate and broken rate.It mainly included the design of grain collecting device,machine vision system and power module.The working process of the combine harvester was analyzed,and a grain collecting device was designed,which was installed under the grain outlet of the combine harvester.The grain samples were collected periodically to obtain relatively static grain samples.Machine vision system mainly included processor,camera,lens,light source selection and lighting scheme design.The embedded processor Jetson TK1 was selected to control grain collection,image processing and computation.Through industrial camera with fixed focus lens,LED light source and forward lighting scheme,the vision system could adapt to harsh field working environment and collect complex grain images.3.The recognition algorithms of impurities and broken rice in complex rice images were studied.MSRCR algorithm was used as the image enhancement method in this paper.The image was segmented by color difference of complete rice,broken rice and impurities in HSV color model,and the recognition results were obtained by using their respective shape features.The 200 rice images collected by the monitoring device were identified,and the recognition results were evaluated at the pixel level.The accuracy and recall of the recognition results are comprehensively measured by F1 value.The recognition rates of stem impurities,branch impurities and broken rice were 86.74%,84.90%,84.68% respectively.The algorithm could effectively identify impurities and broken grains in rice images.4.Software design of on-line monitoring device for impurity rate and broken rate.The software module included grain collection,image acquisition,image processing,display interface,serial port transmission and data storage module.The grain collection program controlled the collection and release of grains by electromagnets.The image acquisition program controlled the camera to collect grains images.The image processing program completed the image processing and the calculation of impurity rate and broken rate.The display interface program displayed real-time grain images,impurity rate,broken rate calculation values and corresponding bar charts through the interface.The serial port transmission module transmitted the data to the upper computer.Data storage module saved the results to SD card for off-line research.5.The performance of the monitoring device for impurity rate and broken rate was verified by experiments.In the indoor test,the average relative errors of impurity rate and broken rate were 9.13% and 8.94%,respectively.In the field experiment,the monitoring device was installed on the combine harvester for on-line monitoring,and the impurity rate and broken rate data were obtained every 4 seconds.The average relative errors of impurity rate and broken rate monitoring were 11.06% and 10.91% respectively,which met the expected effect of device design.