对由光源颜色变化引起的图像色彩偏差,进行了校正,并在YCbCr颜色空间建立了Cb-Cr色度查找表和亮度信息联合的肤色模型,应用预处理技术,去除部分非人脸区域,减少人脸检测的搜索空间,并采用模板匹配方法在人脸候选区域检测人脸.实验表明,该方法能够有效的从复杂环境的彩色图像中检测出左右旋转不超过45°的人脸,且不受人脸表情、尺度和数目的影响,且错误率较低.-Color by the light source caused by the change of image color deviation, a correction, and YCbCr color space established a Cb-Cr chrominance and luminance information look-up table of the color model of the joint application of pre-treatment technology, to remove some non-human face region, Face detection to reduce the search space, using a template matching method in the face candidate region detection of human faces. experiments show that the method can be effective in complex environments from color images detected no more than about 45 ° rotation of the human face, and from Facial Expression, scale and number of impact, and lower error rate.