基于特征库的彩色图像二值化方法

Color image binarization method based on the characteristics library

由于彩色物体图像的三基色存在相关性,三维空间阈值区域呈现复杂形态,因此采用标准形状的阈值区域,将会造成所研究图像阈值范围的扩张,从而降低物料检测和分级的精度。采用基于背景特征库的彩色图像二值化分割方法,可以满足任意形状阈值区间的检测,精度取决于三基色数字采样的位数,且具有快速实时处理方法。此外,对于特征库建立过程中的相关问题,进行了分析和讨论。结合实际工程示例给出该方法实验结果。

Because of the correlation of color image three primary-color signals, and the threshold region of three-dimensional space emerge complex pattern, so it will extend the threshold region of the researched image and reduce the precision of detection and classify if we take standard form of the threshold region. The color image binarization segment method based on background characteristics library is presented, threshold region for all arbitrary shapes can be detected in which the accuracy depend on the bit of the three primary-color digital sample. It also can effectively improve the running speed. In addition, the related questions of the establishment of characteristics library are analyzed and discussed. Finally, combined with practical engineering examples, the experimental results of the method are obtained.