多尺度分形特征在木材内部缺陷检测中的应用

将数字图像处理与木材检测相结合,对获取的木材图像进行图像处理与分析,获取高质量的木材图像是提高木材检测准确性的重要手段。利用微光摄像机作为图像输入传感器,并最终以数字信号的形式储存在计算机当中。应用数字图像处理技术对采集到的原始图像进行预处理,提高图像的质量,便于后续处理。将预处理后的图像分割成若干个子区域,计算出每个子区域的多尺度分形特征值。实验结果显示,图像中背景部分与边缘部分多尺度分形特征值存在明显差别,可以提取奇异性较大的多尺度分形特征值,他们的集合即为木材缺陷的边缘。

Application of Multi-scale Fractal Feature in Wood Interior Defects Detection

High quality wood image is very important in improving the accuracy of wood detection,which can be obtained by combining digital image process and wood detection technology and analyzing the image using appropriate method.The wood image signal is transformed from low light camera into computer.Digital image processing technology is selected to pre-process the original wood image,so that the quality of the image is enhanced and more suitable for later processing.The wood image is segmented into sub areas,and the multi-scale fractal feature values for each sub area are computed.The experimental results show that the values of the multi-scale fractal feature values are significantly different between the background regions and the edge regions in a wood image.Therefore,the edge in the log image can be extracted according to the different values of the multi-scale fractal feature values.