基于结构谱的中厚板表面缺陷识别方法

为克服传统纹理分析的缺陷识别结果易受光照变化和氧化铁皮不利影响的缺点,提出了结构谱纹理分析方法,并将其应用于中厚板表面麻点、夹杂、结疤等缺陷的识别.实验结果表明,结构谱方法具有较好的光照不变性,对麻点、夹杂、结疤等缺陷的识别率要高于灰度共生矩阵、Laws纹理能量、傅里叶功率谱等其他纹理分析方法.

Recognition of surface defects on medium and heavy plates based on structure spectrum

In surface inspection of medium and heavy plates, defects recognition based on texture analysis suffers from mutative illuminations and scales. An illumination invariant texture analysis method named structure spectrum was proposed and applied to recognition of surface defects on medium and heavy plates. Compared with other textural features such as gray level co-occurrence matrix, Laws texture energy, and Fourier power spectrum, higher classification rates were made by structure spectrum for classification of pits, scars and inclusions.