Research On The Cleaning System Of Intelligent Combine Harvester Based On Unmanned Aerial Vehicle Image

The combine harvester is an important part of the agricultural machinery and equipment.The development and research of the intelligent control harvester will provide a solid guarantee for the harvesting of the crops.The harvesting results of the cleaning system for the harvester directly determine the performance of the combine harvester.Therefore,the key to realize intelligent control of harvesting is to develop a fine cleaning device and formulate the corresponding intelligent strategy.This dissertation is supports by the 13th Five-Year National Key Research and Development Program named Intelligent Rice and Wheat Combine Harvester Multi-parameters Fusion Control Strategy.We take the combine harvester's cleaning control system as the test platform,and various sensor data collection is installed in the system.By using the mechanics,sensor technology and intelligent control technology,some intelligent control problem and related technique were studied in the following.First,when the combine harvester is used for a large-scale operations,the obtaining of the characteristic parameters of the machine play an important role during the process of harvesting.The information of the grass-to-valley ration is the most vital among them.The traditional remote sensing technology has achieved success in crop identification,upswing monitoring and yield assessment and so on,however,it has not been applied in the field of crops and valleys for the reason of long acquisition period and high cost.In order to break through the shortcomings of traditional remote sensing technology,we use the unmanned aerial vehicle as a remote sensing platform.Based on those characteristics and the requirements of the target,an algorithm based on image gray layering method to predict crop grass-to-grain ratio is proposed.The algorithm divides the gray value of the mosaic images into several layers according to the difference of the gray value of the images,and acquires the relative magnitudes of the grass-to-valley ratio of its corresponding crop area according to the number of levels.The field data shows that the method has certain reliability.Second,after obtaining the relative magnitudes of the crop grass-to-valley in the target area,a corresponding fuzzy logic control algorithm should be set according to the grass-to-valley ratio distribution.The control system of the combine harvester is belong to a multivariable control system.It is difficult to develop an ideal fuzzy rule base for the harvester cleaning system through the traditional multivariable fuzzy control method.To change the above situation,we propose an incremental fuzzy logic control method.This method formulates rule base mainly by changing the fuzzy rule statement,that is,all the input and output variables are all used as the input part of the fuzzy rule,and the output of the fuzzy rule becomes the increment of the output.The establishment of the incremental fuzzy control model for the combine harvester cleaning system mainly depends on the combination of the experimental data of clearing performance and the experience of the operators.Third,in order to verify the rationality of the incremental fuzzy control model for the cleaning system,we built a complete set of cleaning platform.Based on this platform,we adopt the intelligent fuzzy control method based on unmanned aerial vehicle and traditional manual control method for field application experiments,respectively.The combine harvester were harvested at low,medium and high speeds respectively.The average values of impurities rate and losses rate obtained by incremental fuzzy logic control method were 2.16%and 2.32%respectively.Compared with the data from traditional control methods,they were 2.39%and 2.56%respectively.From the comparison of experimental data that we collected by our manual method,it can be seen that the incremental fuzzy control method based on the unmanned aerial vehicle's imagery is better than the traditional control method,which verifies the rationality of the incremental fuzzy control model.