Dissertation > Excellent graduate degree dissertation topics show

Research on Evaluation of Knitted Fabric Pilling Grade Based on Three-Dimensional Scanning Technoligy

Author: YuShiPing
Tutor: LongRuHai
School: Donghua University
Course: Digital Textile
Keywords: Three-dimensional scanning Eigenvalue analysis Knitted fabric pilling OpenGL simulation
CLC: TP391.41
Type: Master's thesis
Year: 2012
Downloads: 20
Quote: 0
Read: Download Dissertation

Abstract

Knitted fabric fuzz ball is an important indicator to reflect the performance quality of the knitted fabric in the knitting industry, both or trade in textile design, measurement knitted fabric pilling are essential. The traditional approach is to get directly through the observation of the technical staff, when large quantities of detection, it will demonstrate its inherent limitations, not only time-consuming, and the measurement results are also heavily influenced by subjective factors. Improve efficiency in order to improve the traditional methods of measurement, in recent years, some new methods came into being, and the following two ways: first, a two-dimensional image processing method. Pictures of the fabric, and then through a certain method to identify the boundary threshold of wool ball wool ball isolated in order to find the various hair bulb eigenvalue, and evaluation of fabric pilling level; Second, use the indirect method to get the fabric The surface of the three-dimensional coordinates and the surface information thus obtained reflect the knitted fabric pilling characteristic value, based on these characteristic values ??determine the knitted fabric pilling rating. These methods have changed a great extent the deficiencies of the conventional methods, and the subjective method was extended to the objective method. However, the first method, the time to find the boundary threshold by the impact of the fabric color and tissue is relatively large, the second method will to a large extent by the influence of the approaching point errors. This article by Xystum 3D scanner can be direct and accurate access to information of knitted fabric surface knitted fabric surface reconstruction using OpenGL simulation technology, enabling the visualization of the detection process. This paper studies the Xystum three-dimensional scanner surface coordinate information for knitted fabric pilling evaluation introduced Xystum three-dimensional scanner how to obtain the software material surface, and discussed how regularization scan data to achieve better modeling effect, to achieve a separation of the whole process-based of OpenGL simulation technology data processing and information visualization. In the objective processing the scanned information and the acquired pilling Eigenvalues ??process, use a pitch network structured method is corrected to the fluctuations in the scanning spot position caused by the mechanical vibrations and laser scanning, reflection angle offset, and then use two-dimensional wavelet filter denoising algorithm on the height information of the scan process, to remove the influence of noise caused due to mutations in the color and gloss change and hairy fibers and other impurities caused by the reflected light information. Then use the the height average method and the fixed threshold method, etc. obtains the threshold value of the hair bulb was separated, in the VC simulation program using a different color indicates the two pieces of information, and can watch them the separation effect. Obtains the threshold value information separation after four Unicom traversal method can be used to obtain a hair ball area information, thereby obtaining their average height, the maximum height, the projection area, volume, and other information, and then get the entire scan area Hairball the average height and maximum height, the average projected area, a maximum projected area of ??the total area, the average volume, the maximum volume, the total volume of many reflected pilling degree characteristic value. Through many experiments the control of the contrast, and other testing methods, to find the the ball Level correlation between these characteristic values ??Pilling, objective evaluation of the knitted fabric pilling Level. The three-dimensional scanners based Xystum this thesis, data acquisition, data acquisition accuracy very high accuracy micron data obtained information is easy to handle and regularized. Modeling and simulation process The Playwork modeling software as a contrast and a frame of reference, to enhance the efficiency and accuracy of the modeling. Convert software optimized for data, the first to use the scan data into ASC Ⅱ format easy-to-read computer files, then use the VC language of these data is read into the model, using scanning point spacing network structured corrected scanning process offset in the XY plane, through the use of the median filter and the discrete two-dimensional wavelet transform method for processing, in order to effectively reduce impurities and filament fibers as well as machine vibrations caused deviation of the optical data, making the height coordinate data information authenticity of the effectiveness and separation effect is improved. Automatic threshold method and global auto hair bulb separation, this paper height means method, fixed threshold method as well as single-line backward abatement backward abatement automatic threshold method are four ways to get good reference, use different fabrics on different methods the effect is not the same. VC simulation technology for its excellent visibility operations the separation can watch on your computer, so choose the demand threshold method. The trade-off factors and can be added to achieve subjective, objective and effective combination. For objective assessment of pilling is to use the threshold model hair bulb isolated and determined the their Pilling characteristics to value, so the establishment of every scanned image feature vector, and then calculate these vectors to the standard level eigenvectors distance from a certain minimum level, the judgment of the scan knitted fabric pilling level for this level.

Related Dissertations

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net Mobile