Research On The Key Technology Of Branch Recognition And Location Of Intelligent Chinese Wolfberry Picking Robot

Chinese wolfberry is one of the large-scale crops planted in China and an important part of agricultural economic development.It has not only edible value but also medicinal value.At present,the harvesting work is mainly done by manpower,and the picking workers have high labor intensity,high picking costs and low picking efficiency.Therefore,the development of intelligent picking robots is very important.With the rapid development of agricultural robotics and computer technology,it has become possible to manufacture intelligent picking robots.The fruit of Chinese wolfberry is not only small in size but also in large quantity.The single fruit picking efficiency is low.Picking according to the shape of the branches can not only ensure the picking efficiency but also protect the eucalyptus.The picking robot adopts the anthropomorphic design,the visual recognition system and the double-arm structure greatly simulate the working process of the person picking the cockroach,and ensure the smooth picking of the cockroach.With the development of a large number of agricultural picking robots,binocular vision and robotic arm control technology are widely used.The traditional rough picking scheme has been unable to meet the requirements of modern agriculture for picking robots.The demand for intelligent high-precision agricultural picking robots is huge.Vibratory picking mechanisms have been used in the sputum industry,but there are problems with damaged fruits and fruit trees.It can be harvested three times a year,and damage to fruit trees will inevitably lead to problems such as reduced yield and reduced quality.The carding picking device can effectively protect the fruit and fruit trees,but now it is in the hand-held portable stage,and the use process is too dependent on human judgment.The computer vision recognition system can accurately identify and locate the position of the branches,select the position information of the target branches to be picked by the position of the fruit,and use other branches information as obstacle avoidance information.Accurate identification and location of the branches is the key to ensuring successful picking of picking robots.This paper has carried out several researches on the design concept of intelligent picking robot and the key technology of branch recognition.(1)According to the actual growth and picking process,the design idea and working process of the intelligent picking robot are introduced.The intelligent picking robot adopts the anthropomorphic design,and the working process is divided into the branch identification process,the branch grabbing process and the picking process.Intelligent picking robots can maximize the success rate of picking.A description of the design concept and function of the components of the picking robot.Introduce the role and workflow of the various components of the picking process and the picking process.(2)Identify branches in the natural environment.According to the color characteristics of RGB,HIS,HSV,Lab,etc.,select the best color space.The original picture captured by the camera has problems such as noise interference,and the image can be preprocessed by image enhancement and filters to improve the success rate of branch recognition.In the identification of needle branches,it is necessary to identify the local branches and the whole branches,and identify the local branches and the whole branches respectively.Provide the target picking position and obstacle avoidance information for the picking robot.(3)Aiming at the problem of many disturbance factors in the natural environment,the branches are processed by k-means clustering method,Otsu threshold segmentation method and iterative threshold segmentation method to identify the branches.When identifying lychee strips,it is necessary to remove disturbing factors such as fruits,leaves,sky,and land.In particular,the leaves and lychee are similar in color,and the leaves are dense and dense with the growth of lychee.The single method can not effectively remove the interference,and the combination of k-means clustering and Otsu threshold segmentation can effectively segment the litchi.The branching skeleton is extracted by the thin line method,and the spatial position information of the litchi strip is reconstructed by the prior knowledge of the obtained key points of the branch and the branch radius.(4)Aiming at the problem of occlusion of branches by fruits and leaves,this paper puts forward the idea of locating lychee by the location information of the fruit.Because the fruit and leaves are accompanied by the growth of the branches,especially at the end of the branches,the branches are obscured,and the camera cannot effectively read the branch coordinate information.The picking robot needs to determine the position of the picking point at the end of the branch,which can be assisted by the fruit position on the branch.Design experiments and verify the feasibility of assisted positioning experiments.It can provide effective lychee grabbing points to ensure the smooth completion of the picking work.