Research On The Key Algorithm Of Irregular Sheet Soft Object Grasping Process

Leather is widely used in daily life,but its processing technology is extremely complex,which requires a lot of labor,and has high labor intensity and poor working environment.In recent years,due to environmental protection,labor shortage and other factors,the development of China’s leather industry has entered a relatively difficult stage,and leather enterprises have encountered severe employment difficulties.With the application of computer vision technology in the leather machinery industry,the intelligentization of leather equipment has become an important direction for the development of intelligent manufacturing technology.The leather products in the production process belong to irregular sheet soft matter,and the shape is randomly distributed in each process.The existing equipment can not directly realize the grabbing and spreading and stretching and stretching operations of leather,which leads to the fact that the two operations need to be completed manually at this stage.Therefore,it is the key to the intelligentization of leather making equipment to study the method of automatically grasping,spreading and stretching leather.Based on computer vision technology,this paper studies the key algorithms applied in leather grasping process.The main research contents are as follows:(1)Theoretical analysis on key actions of leather processingIn this paper,the sheepskin processing technology is taken as an example.There are 45 processes in total.Among them,manual handling and spreading leather is needed in the transition stage from wet processing to dry processing,and manual selection of stretching points is needed in the drying process of the bandage plate.Aiming at the problem of automatic grasping and spreading in the above leather processing process,a method of grasping and spreading leather alternately with two arms is proposed.Firstly,the leather is initially spread by the method of alternating grasping of two manipulators to find the lowest point;then,according to the color difference between the inner and outer sides of the leather overlapping part,the overlapping part after the initial spreading of the leather is identified.Finally,find the grasping point of the overlapping part to fully unfold the leather.At the same time,aiming at the problem that leather needs to use artificial experience to select the stretching points in the drying process of the bandage plate,a method for selecting the stretching points of leather is proposed.The skeleton of leather is used to obtain six stretching points on both sides of the limbs and head.The remaining stretching points are obtained by the circle intersection method.(2)Research on the algorithm of double-arm stretch spreading leatherAccording to the method of alternately grasping and spreading leather with two arms,the algorithm of grasping point selection in leather spreading is designed.The image is preprocessed by image graying and image enhancement.At the same time,the Gaussian filtering algorithm is used to denoise the leather image,and the obtained gray image is binarized by OTSU method.Finally,the image is segmented by Canny algorithm,and then the effective area of leather image is obtained.The leather image of the overlapping part after the initial expansion can be obtained by adjusting the threshold.The Harris corner detection algorithm is used to obtain the feature points that the manipulator needs to grasp in the effective area of the leather image.Therefore,combined with the above leather grasping and spreading method,a complete detection algorithm for grasping points of irregular sheet soft objects is established.Leather spreading can be realized by alternately grasping feature points with two manipulators.(3)Research on the algorithm of leather stretch point selectionAccording to the method of leather stretch point selection,a circle intersection algorithm based on skeleton extraction is designed.Using the pre-processed leather image in the grab point selection algorithm,the leather skeleton is extracted by the morphological skeleton extraction algorithm,and then the leather skeleton is used to determine the six selected stretching points on both sides of the limbs and head of the sheep in the leather.Taking one of the selected stretch points as the starting point,the intersection of the leather contour and the circle is extracted by the circle intersection algorithm,so as to determine the remaining stretch points.Finally,according to the distance from the required stretching point,the unreasonable stretching point is selected,and the final stretching point of the leather can be obtained after discarding it.(4)Algorithm verification and experimental researchBased on the dual manipulator and industrial camera,the experimental platform of double-arm alternating grasping leather is built,and the simulation and experiment of the leather grasping and spreading algorithm and the stretching point selection algorithm proposed in this paper are carried out.The simulation results show that the Harris algorithm and morphological skeleton extraction algorithm selected in this paper have good practicability.Through experiments,the threshold for identifying the overlapping parts of leather is 0.4,and the best starting point of the circle intersection algorithm is the left or right hind leg.At the same time,the effectiveness of the leather grasping and spreading algorithm and the stretch point selection algorithm proposed in this paper and the feasibility of industrial application are verified.Through the research on the algorithm of double-arm stretching and spreading leather and the selection of leather stretching points,the automation and intelligence of leather grasping and spreading and stretching are realized,which lays a foundation for the subsequent research on the intelligent operation robot of irregular sheet soft objects.