摘要
由于光线分布不均匀或斑块噪音干扰等原因,往往使所要处理的指纹的灰度值分布缺乏均匀性。在指纹特征自动识别提取过程中,造成许多传统的算法在局部出现很大的误差。利用方差和均值特征的自动提取方法,首次对不均匀灰度图像进行自适应分割。然后通过对图像分区域进行不同程度的自适应调整,使具有相同属性的像素单元具有近似的灰度值分布。调整结果的灰度均匀水平与预先指定的调整精度成正比。这种调整不仅提高了图像分割的自适应性,而且进一步扩大了一般阈值算法的应用领域。
The extracted objects with different ranges usually have different di stribution of gray scale value, due to uneven distribution of light and various properties of the crack background of fingerprints. Within the automatic identif ication system for fingerprints, most of all the existed algorithms have led to much error. The adaptive analysis towards images is being conducted by taking us e of the characteristics of variance and mean. Then on the basis of the analysis , the images are segmented adaptively and the uneven density scale is undergoing adaptive adjustment according to the variance and mean of every part of the ima ge. Since the image units with the same density attributes to the direct ratio between gray scale value and preassigned adjusting accuracy. Th erefore it enhances the adaptivity of segmentation and broadens the applied field of the common algorithms for images.
出处
《控制工程》
CSCD
2003年第3期249-251,281,共4页
Control Engineering of China
基金
国家自然科学基金资助项目(70171056
60084003)
黑龙江省博士后流动站和沈阳师范大学科研启动基金资助项目(054 554160)
