| Rape seeding is an important link of rape production.2BFQ-6 precision planter developed by research group has a high degree of functional integration,it mounted on a tractor,they formed rape seeder unit.The rape seeder unit can improve the production efficiency,rape precision seeding,rotary tillage ditch,fertilization cover soil and other links,can be done at once time when it is working in the field.In order to keep the rape seeder working on row precisely,the driver was asked to work with the close coordination of the hands,the brain,the feet and the eyes,it is a challenge to the strength and experience of the drivers.So a study about rape seeder unit working on row based on vision navigation was carried out,it will be helpful to reduce the labor intensity,increase operation efficiency,etc.In this paper,the working path tracking system based on machine vision was designed to make rape seeder unit automatic work on row,specific research contents are summarized as follows:(1)Rape seeder unit vision navigation system which consist of software and hardware was built.Hardware system was made of computer,industrial camera,NI data acquisition card,the power,step-motor,and so on.These function can be implemented through the hardware system,for example,front wheel angle information of rape seeder unit can be obtained,front wheel steering can be controlled electrically,etc.Software system is a MFC interface PC application,it was programmed by VS2012 in Window7 through C++ computer language with open source image processing library OpenCV,and it mainly includes image processing,automatic control and sensor information acquisition module,etc.(2)This paper proposes a kind of vision navigation identification method based on wavelet transform and improving random sample consensus(RANSAC).Firstly,the original image grayed is wavelet transformed to highlight the macro profile of the navigation path.Secondly,using the feature that work area and not work area have a high-contrast in the work images of rape direct seeder,a method which finding the feature points of navigation path is used through ROI sliding windows.Finally,the improved RANSAC with pre-test and post processing correction is used to distinguish inner points and outer points in the set of navigation path feature points,and the navigation path parameters is calculated by least squares in the set of inner points.Field experimental tests show that the algorithm can be stably and accurately detect the navigation path,and the correct rate is 90%,while the processing time for each frame is less than 55 ms.It can provide technical support for rape direct seeder’s vision navigation.(3)A fuzzy controller was designed to path tracking control of rape seeder unit.Firstly,the mathematical relationship between the actual path and the image detection path was derived according to camera pinhole imaging model.Secondly,through combination of two-wheeled Ackerman steering model of rape seeder unit,the change rule of image detection path was analyzed clearly when rape seeder unit is close to or away from the filed sowing path.Finally,based on these rules,fuzzy controller was designed to path tracking control,it can presume the changes of rape seeder’s position through the change information of image path.And,the simulation test result of fuzzy controller proves that this set of strategy is effective in the theory by using MATLAB/ Simulink.(4)A PD controller was designed to control the front wheel steering.Through electronically controlled steering system characteristics test of rape seeder unit,it is found that the steering speed is increased with the increase of PWM signal frequency when the frequency changes from 50 Hz to 900 Hz,and the maximum steering speed is 12.6°/s,so the output range of PD controller was set at 50 Hz to 900 Hz.After site test,controller parameters are determined,scale factor P is 250,differential coefficient D is 80,and control cycle T is 40 ms.The front wheel steering control accuracy is 0.5°,the controller can meet the steering control requirements of rape seeder unit.(5)Road and field experiments were carried out.Pavement test results show: maxmum lateral offset can be controlled at 7.8cm,9.6 cm and 14 cm when the rape seeder unit travels along a straight line at a speed of 0.5m/s,0.8 m/s and 1.0 m/s respectively.Field experiments show: maxmum lateral offset can be controlled at 12 cm,14cm and 28 cm when the rape seeder unit travels along the operation path at a speed of 0.5 m/s,0.8 m/s and 1.0 m/s respectively. |
PDF Full Text Request