Research on fabric wrinkle recovery detection based on an improved KCF algorithm
Fabric wrinkle recovery performance is an important indicator for evaluating the shape retention of fabrics.It reflects the ability of fabrics to spontaneously recover their original shape after wrinkling during use and processing.Good wrinkle recovery performance enables fabrics to quickly recover after folding,transportation and cutting,thereby improving utility.The currently widely used standard method for evaluating fabric wrinkle recovery performance is the wrinkle recovery angle test.This involves folding and pressing fabric samples and then measuring the angle of recovery for the resulting creases to assess the fabric's wrinkle recovery capability.However,this method is complex to operate,relies on manual judgment and measurement for testing,yields results vulnerable to subjective effects,and cannot distinguish or quantify the dynamic evolution of wrinkles during different recovery stages.It can only provide an approximate overall evaluation and fails to accurately describe wrinkle recovery patterns in detail.Therefore,there is a need to develop a new testing method to achieve automated monitoring and quantified parameterization analysis of the entire wrinkle recovery process,in order to more efficiently,precisely and comprehensively evaluate the wrinkle recovery performance of different fabric materials and provide a basis for fabric performance optimization.In view of this,this study proposed a fabric wrinkle recovery detection method based on an improved kernel correlation filter algorithm to achieve automated monitoring and key parameter extraction of the entire fabric wrinkle recovery process.As for the method,a high-speed camera was used to collect and record images throughout the dynamic process of the formation and recovery of fabric wrinkles and a set of detection system was designed to process and analyze the image sequence.In the system,a robustness-enhanced improved KCF algorithm was first applied to track and locate fabric wrinkle areas.This algorithm improved adaptation to texture and shape changes by fusing multiple feature expressions and employed an edge adaptive adjustment algorithm to reduce boundary effects.Next,morphological binarization and skeleton extraction were performed on the wrinkle area of interest to preserve the topological structure while simplifying structural expression.Finally,based on the principles of analytic geometry,polyline fitting was performed on the skeletonized graphical shapes to calculate the vertex angle parameter,thereby obtaining time-series data on the variation of the wrinkle peak angle over time.The method achieved automated monitoring of the entire fabric wrinkle recovery process,avoided subjective judgment,and greatly improved detection efficiency,enabling detection and record of fabric crease recovery at different stages.By comparing the results obtained by this new method with those obtained by the standard wrinkle recovery angle test method,it is found that the two test results have strong correlation,which verifies the effectiveness of this new method and shows that this method can be used as a new technical means to evaluate the crease recovery performance of fabrics more accurately and comprehensively.This study establishes detection technology and analysis methods that lay the foundations for building an intelligent fabric wrinkle recovery evaluation platform going forwards.Future work will progress in the following areas:(i)to continue to improve algorithm robustness to address detection issues under conditions involving light,occlusion,etc.;(ii)to expand the algorithm's scope of application to enable generalized detection across more fabric types;(iii)to collect and annotate large quantities of dynamic wrinkle process data to train deep neural network models for intelligent analysis of fabric wrinkle morphology to support structural design and performance enhancement of fabrics.Progress in the above areas will make testing methods more efficient and precise,promote advances in wrinkle recovery evaluation techniques,and provide new ideas to improve fabric structural design and performance.
万方数据
维普
