Research And Implementation Of The Agricultural Robot Navigation Control System Based On Monocular Vision

Since the presence of high cost,low yield,low efficiency and environmental pollution of traditional agriculture in agricultural production management,high-tech agricultural machinery supporting the development of intelligent represented by information technology,which has become the focus of global agricultural development.Farmland navigation technology is the basis of agricultural mechanization intellectualization and the premise of agricultural intensification planting.So it is very important to study the autonomous navigation technology of agricultural machinery.In this paper,the main research about the agricultural robot navigation control system is as follows:(1)Study on image pretreatment of farmland seedlings,analysis and comparison of different color space model of seedlings image characteristics and gray-scale treatment,on which the image segmentation algorithm of seedlings and farmland background was studied,and then the binary image is obtained by morphological filter to meet the requirements and the image edge detection processing,to provide a more appropriate edge detection image for the subsequent image processing.(2)Study on method of seedlings center line extraction,which based on Hough transform algorithm by transforming the voting of the points to the whole parameter space into voting only to a certain parameter space and designing dynamic adjustment of the ROI region to improve the Hough transform,and then combined with a linear regression analysis method to fit the seedlings line,to complete the seedlings center line extraction.Finally,the processing flow of extraction line centerline of seedling is designed,and the experimental results show that the improved algorithm is feasible and effective.(3)Design the system communication overall solution,through the two modes of automatic navigation and manual control,the agricultural robot navigation control system can effectively solve the abnormal situation in the field work.And also analyzed the communication data exchange flow,developed the communication protocol and interface of the system and realized the system's serial communication and network communication.(4)Design the visual navigation for the agricultural robot navigation control system,analysis and implement the system's hardware components and complete the system's software development and then carry out the visual navigation control experiments of the agricultural robot.Based on the experiments and the data,the visual navigation algorithm for the system is validity and accuracy.Finally do summary analysis for the whole experiments and research.