The Recognition Research Of Grassland Internal Condition Of Mowing Robot

In recent years,artificial intelligence has been widely used in various fields and has received attention largely.Intelligent mowing robots,representative products of artificial intelligence,which can acquire the surroundings and the conditions of grassland through visual sensors,recognize the grassland and obstacles automatically,realize the functions of route planning,self-navigationand autonomous operations.The purpose of this paper is that identify obstacles and boundaries within the grass by adopting suitable machine vision algorithm.This can provide support for walking and obstacle avoidance of the lawn mower,so as to improve the autonomous ability of mowing the lawn.Research content mainly about:(1)Research and recognize internal obstacle in lawn.For different obstacles in the grassland,different methods were used to identify the obstacles.By comparing the identification of obstacles based on color characteristics and texture features,finally,the Mean shift algorithm combines the color and texture features is used to identify obstacles.Experimental results show that this method is simple,butovercoming the defects of single feature method effectively.(2)Inner boundary in lawn segmentation method has been researched.The influence of the unsynchronized length and the angle selection on the texture features is discussed,using the sliding window of filtering to calculate the texture features,in order to saving the time of system,a segmentation method are proposed using texture features and SVM,the initial segmentation was achieved by this method.(3)The inner boundary of grassland is extracted and mapped.By counting the black pixels numbers that left and right in the binary image to judge which grass area is unhandled,then,processing the outside area and the connected area by morphology operation,last,describingand mapping the boundary to the original image by using edge detection operators.