Research On Navigation Path Recognition Algorithm Of Ratooning Rice Harvester Based On Machine Vision

Ratooning rice is a special kind of rice,which can harvest two crops at a time.As a big agricultural country,China has developed the planting technology of ratooning rice to increase the yield of ratooning rice,it is greatly significant for China to develop a powerful agricultural country and to improve the comprehensive benefits of rice planting.At present,domestic research on increasing the yield of ratooning rice by planting high quality seeds with strong regeneration ability,improving the cultivation techniques of ratooning rice,applying fertilizer scientifically,using water scientifically and keeping reasonable planting interval.With the rapid development of machine vision,the application of machine vision technology in the field of agricultural machinery has become more extensive.This paper is devoted to the research of the navigation path recognition algorithm of ratooning rice harvester based on machine vision,the purpose of this paper is to use machine vision technology in the harvester to avoid the wheel crushing on the pile of the first ratooning rice and to reduce the damage of the wheel to the dormant teeth of the stubble of ratooning rice,so as to improve the yield of the second ratooning rice.The research contents are as follows:1.Image acquisition.The images used in this study were collected by the research group in the field of mechanical harvesting experiment of the first-crop rice,and the algorithm involved in this paper was studied.2.Image preprocessing.Because of the influence of light,shadow and crop density,the color information of the images collected in the experimental field is too rich,so the information redundancy of the images should be reduced,the data should be simplified and the efficiency of the algorithm should be improved.3.Path recognition based on traditional image processing algorithm.This paper propose a new algorithm for recognizing the harvester navigation path based on traditional image processing,OpenCV image processing function library,the C + + language is used to complete the grayscale of the image,filter the noise,adopt the big law binary,the morphology processing,the crop and the road contour line extraction and screening,finally,the method of projection is used to extract the center line of the path navigation of ratooning rice field,and the operation and simulation are completed.4.Path recognition of image processing algorithm based on depth learning.Based on the study of traditional image processing,this paper presents a depth-learning-based algorithm for the identification of Harvester navigation path.The U-net semantic segmentation model,the convolutional neural network classification model and the OpenVINO neural network acceleration tool of Intel are used to accelerate the model.5.Research findings.By comparing the results of image processing algorithm based on traditional image processing with those of image processing algorithm based on depth learning,it is concluded that the traditional image processing algorithm has better effect in processing gray image,compared with the traditional image processing algorithms,the depth learning image processing algorithm has more advantages in accuracy and adaptability than the traditional image processing algorithms.