Research On Automatic Variable Swath Mower Based On Machine Vision Technology

At present,the swath of lawn mowers at home and abroad are mostly fixed and cannot adapt to different working environment.In the environment where the space is narrow or the distance between the obstacles is small,the operation can not be completed smoothly,and the situation of leakage cutting appears.After a comprehensive analysis of the technology and obstacle detection methods of the variable amplitude mower at home and abroad,a new type of lawn mower: automatic variable swath mower based on machine vision system was proposed.The main research contents and conclusions are divided into five aspects as follows:(1)The structure design of the automatic variable swath mower.Firstly,on the basis of the original zero-turn radius mower,an auxiliary cutter was added to both sides of the original cutter head.The PLC controlled the servo-electrical push rod to drive the sub-cutter to swing back and forth to achieve from 1.2m-Unimpeded cutting width of 1.8m.The design size and swing speed of the auxiliary cutter head are calculated.And made a detailed introduction to the control system.(2)The design of the PLC program and host computer program.The host computer program mainly uses the C# language and HALCON software to design the mixed programming and the software interface,and used the SerialPort control to realize the serial communication between the PC machine and the MITSUBISHI PLC,which can ensure the accuracy and the real-time of the communication.At the same time,the stability and reliability of the software are verified in the written software.(3)Design of real-time obstacle detection system based on machine vision technology.The detection system is divided into hardware system construction and visual algorithm design.First,the hardware system is selected,and the camera is calibrated to obtain the internal and external parameters of the camera.The algorithm first uses the photos collected by the hardware system for algorithm development.Using HALCON software and HSV color space model to extract obstacle information,then the dynamic threshold segmentation algorithm is used to process the image,and then the morphological operation is carried out.Finally,the minimum bounding rectangle is generated to complete the extraction of trees.Experiments show that the dynamic threshold segmentation method combined with HSV color model can effectively avoid the influence caused by uneven illumination.(4)Distance measurement of obstacles.It is mainly divided into two aspects,one is the measurement of the depth distance between the obstacle and the mower based on the range model of single frame static image and the principle of small hole imaging,and the other is to measure the distance between the obstacles by the interpolation method.The test shows that the method is accurate and the error rate is about 0.25%,which satisfy the test requirements.(5)Using C# and HALCON software for mixed programming,developed a machine vision measurement software,and a brief introduction to the use of the software and interface.The experiments were carried out in combination with a variable-width lawn mower.The results showed that the designed structure was stable in operation and strong in stability,meeting the design requirements;the software responded quickly and satisfied the requirements of real-time performance.