针对铁磁材料表面裂纹缺陷，采用新型涡流热成像方法进行了检测。分析了感应涡流激励引起裂纹处生热的机理，设计了基于串联谐振的涡流激励装置，构建了高频低功率激励的涡流热成像检测试验系统，利用该系统对铁磁材料试件进行了检测。针对裂纹图像序列存在的“横向模糊”效应，采用高通滤波和锐化等方法对原始热图进行了增强处理，基于直方图双峰法、迭代法和数学形态学的分水岭等方法对裂纹图像进行了分割，实现了裂纹长度参数的定量识别。结果表明：涡流热成像能够快速、高效、直观地检测出铁磁材料的裂纹缺陷，当裂纹长度为5 mm时，误差不超过15%。

For surface crack defects of ferromagnetic materials, the late-model eddy current thermography method was used for testing research. The mechanism of heating in crack caused by eddy current excitation was analyzed, the eddy current excitation device based on series resonance was designed, and the eddy current thermography testing system which was in high frequency and low power was constructed. Through the designed test system, ferromagnetic materials were tested. In view of the effect of "transverse fuzziness" in the image sequence, the enhancement method of high-pass filtering and sharpening on the original heating image, the histogram twin peaks method, the iterative method and the watershed method based on mathematical morphology for the image segmentation were adopted, and the quantitative identification of crack length was finally realized. Experimental results have shown that eddy current thermography can realize a quick, efficient and intuitive detection on crack defects of the specimens, and when the length of crack is 5mm, the error is less than 15 percent.