脆性固体破坏表面裂纹的自动识别和处理

用长距离显微镜和摄象系统获取的表面细观裂纹图像，由于光线等因素的影响，常模糊不清，且夹杂大量的伪劣纹。为了得到与宏观变形响应同步的表面裂纹的定量描 ，必须对图像进行技术处理，重置像素点的像素值，去除信 。

The Auto-Detection and Auto-Processing of Crack Images on the Surface of Brittle Solids

The meso crack images on the specimen surface, acquired by long microscope connected with CCD, are often obscure because of light and others. At the same time, many false cracks are often included. In order to describe surface cracks synchronous with macro deformation in quantity, the images of crack must be processed and false cracks must be deleted. The two steps should be included: One is to extract cracks according to the difference of gray value, in which false cracks are also included; the other is to delete false cracks.