摘要
在轮胎X射线检测图像中,钢丝帘线的不同排列方式构成不同的纹理图像。Gabor函数可提取图像在频域不同尺度和方向的特征,因此采用基于Gabor和Log-Gabor小波对轮胎X射线图像进行特征提取的方法,并用K-means方法进行图像聚类,实现了轮胎不同纹理图像的分割;采用Log-Gabor函数提取不同方向的纹理特征,并对提取结果进行比较,可检测带束层钢丝帘线疏线或顺线缺陷。
In the X - ray tire images, the different arrays of steel cords form different texture images. Using Gabor function, the features of images in different frequency domains and directions can be extract. Therefore features of X - ray tire images are extracted based on Log - Gabor and Gabor wavelet, then im- age segmentations are achieved with K -means clustering; Texture features are extracted in different direc- tions by Log -Gabor function, then the defects of rare cords or missing cords in one direction can be detec- ted by comparing the extraction results.
出处
《机械与电子》
2016年第4期59-61,65,共4页
Machinery & Electronics
关键词
轮胎
GABOR变换
图像分割
缺陷识别
X - ray tire image
Gabor transform
image segmentation
defect recognition
