Research On The Vector Technique Of Embroidery Pattern Based On Machine Vision

Quilting processing is a technology which is closely linked with the life, the technology will trace embroidery pattern outline for machining tracks, complete the relatively fixed between the outer fabric and filling material. Before quilting processing need to plate, plate making is the purpose of generating nc code. At present, the common plate making software is artificial to describe vector graphics of the pattern outline, however, the artificial way of drawing, low working efficiency, and low accuracy, difficult to achieve high quality plate making. Therefore, in the face of complex trace of embroidered pattern, how to quickly implement contour vector, is of great significance.In this paper, combining digital image processing technology and machine vision, studied a suitable vector quantization technology for embroidery pattern, the technology is mainly through figure collection, contour extraction and contour vector three steps. For collection and measurement of patterns, the paper analyses the camera imaging model, including linear model and nonlinear model, and focuses on the two-step calibration method based on plane. Before the contour extraction, image preprocessing to improve the quality of the image. In terms of contour extraction, studied the contour extraction algorithm based on mathematical morphology and edge detection, for a variety of algorithms are analyzed in the experiment, put forward in view of the figure of a single, small size and pattern complex, large size contour extraction algorithm of two kinds of different situations.In vector quantization of pattern outline, basically have contour tracing, contour optimization and vector quantization three key steps. In the contour tracking, mainly using the Freeman chain code tracking algorithm for contour tracing, an orderly sequence of contour points are stored in data chain table of the program. In outline optimization, and puts forward appropriate contour optimization algorithms to solve the singularity and redundant details, this method proved ideal for optimal results. In vector quantization, is put forward in this paper the vector quantization algorithm based on corner detection, first to corner detection of outline design, the whole contour is divided into several segments through corner, and then to straight line fitting contour segments. Experiments show that this algorithm has the following advantages:target characteristics intact, high precision and high efficiency. Based on the theoretical study, the use of Visual c++6.0and OpenCV software to achieve embroidery pattern vector software development, using its complete picture of typical embroidery pattern vector. Through the simulation machining experiment and field experiment proved that high working efficiency of the system, to meet the accuracy requirement of engineering application and ideally quilted effect.