Fabric Spear Flocking Trajectory Planning System Research Based On Image Identification

With the rapid development of economy and improvement of people's living standards,the manual fabric such as carpet and embroidery has been widely used in households,hotels,exhibition halls,companies and other places,which makes it of high demand and a broad market prospect.The production process of handmade carpet mainly contains pattern design,color,yarn dyeing,print the cloth,hand weaving,flat mat,wash the carpet,cast cut,etc.Most of the working procedure is operated entirely by hand.As we all know,manual work requires high labor intensity and is of low efficiency which compromises economic benefit of enterprises.In order to improve the work efficiency,the carpet quality as well as the precise control of colored yarn dosage,more and more companies began to research the use of the spear flocking robot to replace manual work.And the key technology to realize intelligent handmade carpet weaving is to extract and optimize the robot flocking trajectory automatically according to customer's designs.To solve this problem,this paper studied a fabric spear flocking trajectory planning system based on image identification.The main research work and achievements are summarized as follows:(1)The overall scheme of the fabric spear flocking trajectory planning system based on image recognition has been proposed,the system mainly includes the color clustering module and flocking trajectory planning module based on clustering image,and flow chart of the whole system is given.(2)In view of the color image clustering problem,the quantization algorithm based on both color palette and color difference minimum rules were analyzed,then,an adaptive k-means clustering algorithm was proposed.The Euclidean distance was improved using HSI color space which accords with the characteristics of human visual perception and solved the difficult problem of needing to gain the clustering number according to the prior knowledge;the accuracy of color clustering has been increased at the same time.(3)In order to extract flocking trajectory of the clustering image,all kinds of edge detection and contour tracking algorithm principle have been analyzed,closed hook side path of the color piece image was generated using Canny operator and contour tracking algorithm based on chain code.On the bases of color piece circuitous filling algorithm based on line scan performance,a closed path screw filling algorithm based on heuristic search was proposed which reduced the mechanical shock and wear of parts,the flocked fabric quality was also improved significantly.(4)With the enterprise's actual orders as a case,the whole process has been tested,including color clustering,flocking trajectory extraction and NC code generation.Compared with traditional manual mode,the flocking robot can realize a more evenly needle gauge and stitch,the carpet surface quality and the production efficiency has also been improved significantly.