The Application Of Visual Servo In Separation Of The Cotton Seedlings

In recent years, Visual servo technology is developing rapdily, which have made more application results in all walks of life. Machine vision give robots the visual ability to allow the robot got the flexible jib skills, which has great significance for agricultural machinery’s none-damaged homework. This is one of the important research direction for agricultural machinery. At present due to various reasons restriction, visual servo system has less application in agricultural machinery. In this article, the Welfare industry research special (nyhyzx-005) whose name is’studying of automatic manipulator type transplanter for bare-rooted seedling’, is combined to research the application of the visual servo in cotton seedling separation. This article mainly include the following respects:1、it is first discussed that the status of the development of the visual servo system at home and abroad and the application of visual servo in the agricultural machinery, and then the mode of cotton seedling separation is analysed.2、Based on the laboratory existing experimental conditions, the visual servo platform is built. An inverse-kinematics of the servo system is developed and solved. The camera calibration which is a process to acquire camera’s intrinsic and extrinsic parameters including radial distortion is accomplished based on the Zhang Youzheng plane calibration method.3、Two kinds of image processing method are offered and used for processing single seedlings and many seedlings.For single seedlings, the position near the fist cotyledons in the cotton seedlings stem is set to separation position. The color of this section and the leaves is green. This section will be a key to highlight after grayscaling the image of cotton seedlings based on Super green method. The image of cotton is split by the best dividing threshold which is got from Ostu. And for many seedlings, cotton seedling characteristics is difficult to characterize based on the Spur green method because of the cotton seedling interlacing branches.Through the study, the stem is put forward as separating position and the segmentation of the cotton seedling image is achieved based on the color image segmentation. In this method, the grayscale of cotton image is accomplished using the H of the HSV color space,and one method which based on the minimum threshold and envelope of the histogram is proposed to splitting the image of cotton seedling. And the result of this method is better than others.4、The 9×9 weighted template is used to match the thinning image for the single cotton seedling and the method of prevent misjudgment and removing redundant point is designed to make sure the characteristic point is recognized accurately Identification method based on the geometrical information is applied for many cotton seedlings. The cotton seedlings is recognized through the area and length-width ratio of the image block in the cotton seedling’s binary image. Then the barycentric coordinates of the seedling stem is achieved and used to separating.5、The human-computer interface of the visual servo system is compiled through the GUI function of the Matlab.The human-computer interface include the following four interfaces:initial,proper functioning, parameter calibration and system testing.The human-computer interface makes the operation of the experimental platform convenience and high-efficiency, and the system testing interface have high reference value for diagnosis of faults in system.