Research On Establishing Work Area Of Mowing Robot

With the development of social economy,urban greening has become more and more important to the government departments of all over the world.The area of lawn is growing,and the situation of lawn is becoming more and more complex.The traditional lawn mower is given priority to with manual operation,which usually needs to consume much time and labor.In order to reduce the cost of grass cutting,intelligent robot mowers take the place of the traditional manual mower as a trend in recent years.Collecting and analyzing lawn information through camera equipment is a core problems of the intelligent robot mowers.It is also the basis of intelligent robot mowers autonomous navigation.The main research contents have the following three aspects:1.In the absence of heavy fog,the collected images are reprocessing,there is a lot of noise during filming and transmission of the equipment.This article first analyzes the noise that grassland pictures may produce According to the characteristics of Gaussian distribution,with a small amount of salt and pepper noise,I improved Gaussian filter based on fuzzy decision is designed to de-noise grassland images.2.Due to the outdoor operations of mowing robots,the fuzzy quality caused by heavy fog and uneven illumination is an urgent problem to be solved.In order to enhance the contrast between grassland and background in grassland images,this paper designs an image dehazing algorithm based on the Retinex theory under the TV model.Through the color space transformation,color restoration and correction,the ability to remove fog is improved.The grassland image edge extraction work provides thebasis.3.The extraction of the edge of the grassland is convenient for the establishment of the working area of the intelligent lawn mower robots.Firstly,using the traditional threshold segmentation method to segment the edge of grassland and to find that the valid marginal information can not be find in many pictures.Therefore this paper introduce a variable structural element to fill and refine the grassland images after binaryzation.Finally adopted the least square method based on a fixed threshold value toextract themarginal information and determine theworkspaceof the robot mowing.