The Research Of The Vectorized Algorithm On The Engineering Drawings

This paper mainly describes the Binary-conversion of image algorithms, the methods to filter out the noise and the vectorized algorithms. The main work is as follows:1) Firstly, a research on the Binary-conversion of image algorithms is done. Such as the global binary method, the local binary method, and the automatic-adaptive multi-threshold method, and so on. Some experiment results are shown.2) Secondly, it is described the methods which can filter out the noise on the edge of the line structure. A method of morphological filter with multiple structuring elements is presented and a model of multiple structuring elements is designed which could have better results.3) Thirdly, the comparison and analysis among the classic vectorized algorithms are presented. Such as the thinning method, morphological method, contour tracking method, adaptive window method and the global vectorized method. Then, a new global recognized algorithm is introduced in this paper, which is based on the characters of the image's contour line direction. This algorithm stands in the macroscopically view point to extract the characteristic information of the image. This method grasps the main topology structures of the image. It also has more efficiency and better result compared with the traditional global recognized algorithm, especially in the processing of the inflexions and the crossings.