Visual System Research On Apple Harvesting Robot

The degree of agricultural modernization can be manifested by the level of agricultural mechanization,and and the agricultural production efficiency will be greatly improved after the introduction of robot technology,and the application scale and depth of robot technology also represent the overall level of agricultural mechanization of a country.There are many problems facing our countries,such as aging of population,and aging of population and so on,and there is great and urgent needs for the robot to replace the people to engage in the strenuous picking operations.The fruit and vegetable picking robot was quite complicated,and it is a multidisciplinary field.In this paper,the current situation of fruit and vegetable picking robot at home and broad,with the emphasis on the vision system of robot.1.Researches have been made on the current situation of studying of vision system of fruit and vegetable picking robot,and the position of vision system in the fruit and vegetable picking robot has been elaborated.Combining the researches at home and broad,the application of artificial intelligence in picking robot is discussed.2.Through the R-G model for chromatic aberration,by means of improved Otus maximum interclass variance technology,the target area of the apple image is extracted,and then the close fruits are separated by circular Hough transform,and the necessary experimental verification is carried out at the same time.When the relevant features are extracted,the two-step extraction method is adopted for detailed elaboration.At the same time,based on this algorithm,the preliminary software can be prepared after the reasonable improvement.With the software,the target can be effectively identified from fruits and vegetables of different colors,and the feature information can be extracted.3.Based on the software of Open Source Computer Vision Library,the 3D correction,match and distance measurement have been conducted for the binocular stereoscopic vision.In the measurement of binocular stereoscopic vision at home,the researchers are limited to the theoretical analysis of structural parameters such as baseline distance B,optical axis and Z axis angle ? and so on,and it is difficult to measure during experiment.Aiming at this problem,the author has made creative move and established a sophisticated comprehensive experiment platform for the error analysis of binocular visual system.By combining the experiment platform,the author further proposed a structural parameter model of binocular visual system represented by rotation radius r and rotation angle ?.In this experiment platform,the structural parameters can be accurately set and read,reducing the error caused by manual operation and improving the experimental efficiency.The influence of structural parameters on the system error is analyzed.Through the theoretical analysis and experiment,it is proved that the experimental platform and the structural parameter model of binocular visual system is of the application value.