Recognition Technology And Experimental Research Of Orchard Driving Area Based On Machine Vision

This study is aimed at standardized apple orchard,in view of the labor shortage in orchard operation and the impact of interference factors on the recognition effect of the driving area,this paper proposed the recognition methods for the driving area between rows of fruit trees and the unstructured road in the orchard,as well as the agricultural robot classification of work scene and the boundary tracking method for the driving area between rows of fruit trees.The method in this paper provides research ideas for the recognition of the driving area of agricultural robots in most standardized orchard complex environments.The specific research contents are as follows:(1)In order to solve the problem that the recognition of the driving area between rows of fruit trees is easily affected by factors such as shadows and weeds,a new algorithm based on double robust regression was proposed for the recognition of the driving area between rows of fruit trees in a complex environment with the sky as the background.The tree crown and the background sky were separated by the blue component(B component),and the Otsu algorithm was improved to achieve a better effect of segmentation.After morphological processing,according to the regularity of tree top distribution,dynamic threshold was used to find “V-shaped” region of interest and extract feature points.After the interference points were eliminated by Theil-Sen robustness regression,the straight line at the tree top was fitted by random sample consensus(RANSAC)algorithm,the slope of the straight line at the edge of the driving area was obtained through the slope transformation relationship,and the key point coordinates were obtained using the information of the feature points after elimination and the threshold elimination.Taking the slope as the constraint condition,the linear equation of the edge of the driving area was obtained by substituting the key points.The least square method was used to fit the data for realizing the recognition of the driving area.The experimental results showed that compared with Theil-Sen algorithm and RANSAC algorithm,the average deviation angle of the double robustness regression algorithm in this paper was reduced by 8.28% and9.88%,the standard deviation was reduced by 6.25% and 22.89%,and the accuracy was improved by 4.64% and 10.49%.(2)Aiming at the problems that orchard unstructured roads have no obvious boundary and there are shadows,soil and sand interference at the edge of roads,a method of orchard unstructured road recognition based on feature fusion was proposed.This paper proposed a dynamic region of interest(ROI)extraction method based on the combination of filtering and gradient statistics was proposed to select the ROI of the S component of the HSV color space.The maximum value method was used to merge the color features with the S component mask for multidirectional texture features for binarization and noise reduction.The feature points were found according to the abrupt features of road edges,and a two-level pseudo feature points elimination method based on the dual constraints of distance and position was proposed.To better fit the irregular edges of unstructured road,the method of segmentation cubic spline interpolation was introduced to fit the road edges to realized road recognition.The experimental results showed that compared with the S-component and texture images,the average longitudinal deviation was reduced by44.03% and 96.58%,and the average deviation rate was reduced by 45.45% and 97.00%under multiple working conditions.The mean value of average deviations of least squares method,random sample consensus method(RANSAC)and segmentation cubic spline interpolation method for fitting edges are 2.64 pixels,3.16 pixels and 0.66 pixels respectively,the mean value of average deviation rates are 1.02%,1.21% and 0.26%respectively,and the average standard deviations of deviation rate are 0.23%,0.31% and0.09% respectively.The mean value of average deviation,mean value of average deviation rate and average standard deviation of deviation rate of the algorithm in this paper are the minimum,which indicate that the fitting method in this paper has higher fitting accuracy and better stability.Under multiple working conditions,the average processing time of a single image of this algorithm is 89.9 millisecond,which meets the real-time requirements of agricultural robots in the process of operation.(3)According to the different recognition algorithms of the driving area between rows of fruit trees in different periods and unstructured road,this paper proposed to use convolution neural network for scene recognition,so that agricultural robots can correctly select driving area recognition algorithms according to the working environment.Select Kalman filter to track the boundaries of the driving area between rows of fruit trees to improve the stability and real-time performance of the recognition algorithm.By analyzed the advantages and disadvantages of each network model,this paper finally selected VGG16 model to classify three kinds of work scene and analyzed the test effect of this model.The experimental results showed that the accuracy rate of the model classification in this paper can reach 96.85%,and each category had good robustness.It can achieve better scene recognition in the actual work environment,which lays a good foundation for the subsequent correct selection of the driving area recognition algorithm.