CAJ | 학술논문

针对自动色彩均衡( ACE)算法不能很好地保持图像原始色彩信息，并且算法复杂度高较难满足实时性应用的缺点，提出双选像素点的局部自适应ACE算法。将获得的图像变换到YCbCr色彩空间中，有利于保持图像的原始色彩信息，利用亮度图像的梯度信息与图像局部的均值方差信息进行双层像素点选择，降低算法的运算复杂度，采用局部自适应滤波调整ACE算法中的亮度控制函数，增强图像的局部对比度。实验结果表明，与传统的ACE算法相比，色彩恢复多尺度Retinex算法MSRCR的视觉效果更好，运行速度提高约150倍，计算复杂度得到了明显改进。
침대자동색채균형( ACE)산법불능흔호지보지도상원시색채신식，병차산법복잡도고교난만족실시성응용적결점，제출쌍선상소점적국부자괄응ACE산법。장획득적도상변환도YCbCr색채공간중，유리우보지도상적원시색채신식，이용량도도상적제도신식여도상국부적균치방차신식진행쌍층상소점선택，강저산법적운산복잡도，채용국부자괄응려파조정ACE산법중적량도공제함수，증강도상적국부대비도。실험결과표명，여전통적ACE산법상비，색채회복다척도Retinex산법MSRCR적시각효과경호，운행속도제고약150배，계산복잡도득도료명현개진。
An Automatic Color Equalization( ACE) algorithm based on double selective pixels and local adaptability is proposed in this paper to solve the problem that traditional ACE algorithm can’ t keep original color information,and it is difficult to meet the real-time application for its highly complexity. The proposed algorithm transforms the obtained image into the YCbCr color space to hold the original color information, and reduces the computation complexity of the algorithm through selecting pixels with the gradient information of intensity image and the local mean-variance information. The local contrast is enhanced because of using local adaptive filtering to adjust the brightness control function of ACE algorithm. Experimental results show that the improved algorithm gets a better result which satisfies the human visual properties compares with Multi-scale Retinex with Color Restoration ( MSRCR ) algorithm and traditional ACE algorithm. Computational complexity is greatly improved as the algorithm runs faster than conventional ACE algorithm by about 150 times.