| Sorting of fabric is essential to the Textile industry,the current fabric sorting work is still done by the manual,there are many drawbacks.For example,the efficiency of manual sorting fabrics is very low;Workshop environment inevitably produce dust,noise and other factors to cause some harm to the body of workers;The increase of labor cost leads to the increase of production cost,and the benefit of the enterprise is reduced.With the advent of the INDUSTRIE 4.0,the application of the binocular vision industrial robot is more and more extensive.At present,the main application of the vision robot are involved in automobile manufacturing,electronic engineering,tobacco,food,medicine and other fields.The textile industry has also introduced binocular vision industrial robot to the production process,to speed up the process of textile industry automation.In order to solve the above problems,the fabric of the sorting system based on vision robot is proposed in this paper,which is 20 times the efficiency of manual sorting,and this system avoid environmental hazards caused by production process,improve the efficiency of enterprises,and it has important significance to the textile industry sorting.The main contents are:(1)Build the hardware platform of the system and the design scheme of the image acquisition,the selection of the camera,the selection of the lens,the design of the lighting system and the parameters of the image acquisition card are discussed;(2)Using the color look-up table method(CLUT)to classify the fabric color,color classification not only can achieve a variety of solid color fabrics,and the similar color fabric can also be distinguished.In this method,the idea of linear classifier in pattern recognition is applied to the establishment of mapping relation in color look-up table,and the RGB color space is converted to HSI color space to improve the accuracy of color recognition;(3)Fabric image segmentation using threshold image segmentation algorithm,and get the edge of the segmented fabric image.The segmentation method based on threshold by selecting appropriate threshold,the threshold value is chosen according to the difference between the gray value of gray level of segmented fabric and background.finally,the image segmentation is realized by dividing the pixels into different gray levels by threshold;(4)The point of Robot grasping fabric is set to the center point of the maximum inscribed rectangle of fabric.The method is based on traversing center diffusion.The idea of the algorithm is get the center point of the minimum bounding rectangle,and then moving the center point by the rules which has been set.finally,traversing all the center point,and calculate the rectangular area corresponding to the maximum inscribed rectangle,get the maximum inscribed rectangle,calculate the center point;(5)The principle and function of camera calibration are discussed,and the 2D image coordinate is transformed to 3D world coordinate,and the robot is provided with the coordinate position of the grasping fabric,and complete camera calibration experiment;(6)Introduces the design framework of the PC software of fabric grasping system,which combined with HALCON and VC++,and implement the image acquisition and processing module of the system,and by calling the external dynamic link library functions to achieve the preparation of other modules,finished fabric grasping action.This paper realizes the goal of binocular vision robot grasping fabric,and discusses each method in detail.The system can realize the fabric sorting,the whole process is simple,it is proved that the system has a certain application value in the production of textile industry. |