为了实现工业CT切片图像内外轮廓的准确、快速判别, 提出了工业CT切片图像内外轮廓的自适应判别方法。该方法能根据轮廓的凸凹性特征自动选取相应的内外轮廓判别方法: 首先对单像素宽封闭轮廓的凸凹性进行判断, 然后对存在凹点的切片轮廓采用射线法进行内外轮廓判别, 不存在凹点的切片轮廓则采用坐标极值法进行判别。实验结果表明, 该方法提高了自动判别程度, 扩大了轮廓判别的适用范围, 提高了判别质量和效率。在面向快速原型制造的具有复杂内部结构的大批量内外轮廓判别时, 该方法具有优势。

A self-adaptive identification method is proposed for realizing more accurate and efficient judgment about the inner and outer contours of industrial computed tomography (CT) slice images. The convexity-concavity of the single-pixel-wide closed contour is identified with angle method at first. Then, contours with concave vertices are distinguished to be inner or outer contours with ray method, and contours without concave vertices are distinguished with extreme coordinate value method. The method was chosen to automatically distinguish contours by means of identifying the convexity and concavity of the contours. Thus, the disadvantages of single distinguishing methods, such as ray method’s time-consuming and extreme coordinate method’s fallibility, can be avoided. The experiments prove the adaptability, efficiency, and accuracy of the self-adaptive method.