在分析了多种类型的图象二值化方法在矿图扫描图象分割中存在的问题后，依据图象的集合运算原理，提出了一种兼顾图象整体灰度水平及象元邻域特征的二值化方法——多阈值动态自适应二值化方法。该方法可克服矿图扫描图象二值化过程中出现的线划粘连或断线现象，且方法简单、容易实现。

Image binarization is one of the key steps in scanning digitizing of mining maps. Binarization result directly affects the efficiency and precision of consequent vector reconstruction. After the analysis of various kinds of image binarization algorithms in the processing of scanned mining maps, we bring forward a new image binarization method based on set operation of binary images, which is named as Dynamic Adaptive Binarization with Multi Thresholds. This approach can take into account the gray levels of the whole frame and the neighborhood feature of the pixels simultaneously. With this method, Image binarization is easy and fast , and it can get rid of adhesion and rupture of the lines in the scanned maps.