中文摘要：

针对课堂人工查勤效率较低的情况,提出一种利用纹理和颜色特征的教室人数统计算法。首先提取HSV颜色和LBP纹理特征建立人体头部模型,而后根据头部颜色特点进行图像二值化获取候选感兴趣区域,通过与模型进行匹配得到人体头部区域最终实现人数检测与统计。通过实验说明该算法在教室背景下能较好、准确地实现人数统计。

英文摘要：

Proposes a people counting algorithm in classroom based on HSV and LBP feature is to improve the poor checking presence in class. First- ly, extracts the HSV color features and LBP texture features to build the people head model. Secondly, binarizes the image according to the people＇s head color, and gets the region-of-interest （ROI）. Finally, matches the ROIs with the head model to detect the people in the class- room, and gets the right people number. The experiments demonstrate that the algorithm proposed can count the people in the classroom with high detection rate.