The On-line Measurement System Of Garment Dimension Based On Image Processing

In the process of garment production,quality detection has become the focus of attention in the industry,garment size measurement is very important in the process of garment quality testing.However,the traditional method of measuring the size of clothing,will inevitably appear low efficiency,poor accuracy and other issues,Because of the large amount of clothing,will face the detection of high fatigue,slow detection speed,high detection error rate and high detection costs.Therefore,a method of automatic measurement of clothing size is particularly important,this topic is the machine vision image processing technology is applied to clothing size online measurement field,put forward a kind of high accuracy,fast detection speed and high stability of the clothing size measurement system,which plays an important role in reducing the cost of clothing size measurement for the garment industry.The two aspects of clothing size measurement system of system hardware platform and system software is designed in this toptic.The hardware part of size measurement system is mainly composed of a light source,camera,lens,color plate,computer.Based on the stability and detection precision of the dimension measurement system,the hardware parts are selected reasonably.The software part of garment size measurement system is designed from collecting clothing image,image preprocessing algorithm design,clothing image corner detection algorithm design,clothing size measurement method and human-computer interaction interface design,the image acquisition is based on high performance industrial camera and fixed focal length lens to improve the real-time performance of image acquisition;Image preprocessing includes removing image noise,blurring enhancement,smoothing operation and edge detection;Feature key corner extraction of garment image by corner detection algorithm,aimed at the problem of the traditional corner detection algorithm in high false detection rate,low stability,slow detection speed and difficlut setting threshold,the introduction of adaptive threshold in the study of the theory,implementation of adaptive threshold corner detection algorithm,while the introduction of gradient template theory,multi-scale space feature and Gauss convolution theory are introduced to improve the angle point detection algorithm detection speed;Based on analysising the calibration positioning of the key features of the extracted corner,combined with the size estimates of clothing size measurement,finally obtained the true clothing size,and analysised the measurement of the standard error and repeatability.The garment size measurement system is developed on Visual Studio 2010 environment of the C++ programming,using the integrated environment,the use of clothing size measurement interface is desgined by C# programming,to achieve the measurement of the size of the clothing parameters.Finally,according to the errors of thegarment size measurement result,the reasons of possible errors are analyzed,and the method of reduce the measurement errors is put forward.Through the analysis of the experimental results and the error precision,size measurement system of this research can achieve the standard of garment industry measurement clothing size,the garment size measurement system is applied to the modern industrial production needs to perfected part and improved method is put forward.