Research Of Machine-Vision-based Control System For Weft-Straightener

With the national development strategy of "made in China 2025" put forward by our country,China’s manufacturing began to transform and upgrade to China’s intelligent manufacturing,and the printing and dyeing industry also began to follow the development pace of national intelligent manufacturing,accelerating the upgrading and transformation of enterprise production equipment.At present,most of the printing and dyeing industry uses photoelectric weft-straightener,which has the disadvantages of poor adaptability and low precision for the weft detection of thick,sparse mesh and twill fabrics,which seriously hinder the development of the printing and dyeing industry to high-quality textiles.Machine vision weft skew detection has the advantages of high precision,wide adaptability and high speed,which has become the direction of intelligent upgrading of weft-straightener.Based on the actual needs of a textile equipment company,this paper designs the solution of the weft-straightener control system based on machine vision,and selects the software and hardware involved in the system to build the system’s software and hardware platform.Fabric weft skew detection has the characteristics of wide-format and motion capture,etc.It adopts the method of sending trigger signal to multiple Gig E area-array cameras by means of action command,and capture local fabric images in parallel,and passes the captured multi-channel local fabric images through Gigabit Ethernet switch transmit image data to the industrial computer.According to the characteristics of fabric weft,a weft detection algorithm combining Sobel edge direction detection and progressive probabilistic Hough transform is proposed.The algorithm first uses linear point operation and histogram equalization to enhance the contrast of fabric image.Secondly,it uses fast Fourier transform and frequency domain combined filtering to highlight the yarn texture of fabric image.Again uses Sobel edge direction detection and threshold segmentation method,extract the weft pattern region.Finally,morphological skeletonization and progressive probabilistic Hough transform are used to detect the approximate line of weft and its weft skew angles.The test results show that the algorithm has the advantages of wide adaptability,high precision and high speed.Aiming at the problem of weft skew fusion of local fabric images,a multi-channel vision weft skew fusion algorithm is proposed.According to the tangent value of fabric weft skew angle at each local position,the algorithm establishes a coordinate system for the whole width fabric,uses the least square method to fit the weft deformation curve,and calculates the weft deformation value,which provides an accurate basis for the adjustment of the straightening roller of the weft machine.On the basis of the above algorithm,an application software for the control system of machine vision weft-straightener is developed.The industrial computer runs the control system software,and the PLC controller controls the straightening device.The system software uses the.NET development platform,C# as the programming language,and SQLite database to realize the functions of system user login,system parameter setting,detection image display,weft deformation data curve drawing,manual/automatic control straightening device,historical data query,etc.The test of the control system shows that it runs reliably,all functions are operating normally,and meets the design requirements.