炭素材料X射线图像的缺陷区域提取

针对炭素材料X射线检测图像的特点，对其缺陷提取技术进行了研究，设计了目标边界提取算法和基于小波变换的图像增强算法。鉴于图像边缘特征易受噪声影响，而区域特征受噪声影响要小得多，提出利用数学形态学和最佳阈值分割相结合的方法来实现缺陷区域提取。实验结果表明，该法受噪声影响较小，能较好地实现缺陷的自动提取与分割，为进一步的缺陷特征参数的提取与选择奠定了良好的基础。

DEFECT REGION EXTRACTION IN X-RAY IMAGES OF CARBON MATERIALS

Regarding the characteristic of X-ray inspection images of carbon materials, defect extraction techniques are studied, for this purpose the target boundary extraction algorithm and the image enhancement algorithm based on wavelet are developed. Because the images edge can be affected by noise more easLly than the region feature, defect region - detection is proposed by combining mathematics morphology and optimal threshold method. The experimental resuhs indicate that the defect extraction method can achieve defect extraction and segmentation automatically, which will lay a good foundation for the extraction of defect feature parameters.