中文摘要：

针对煤矿井下照度低、粉尘大、强光干扰多而导致监控视频图像模糊、容易出现目标检测失误等问题,提出了一种煤矿井下监控视频图像预处理方法。引入直方图均衡化算法将灰度值从集中在某段灰度级区域的非均匀分布状态变换为均匀分布状态,扩展像素灰度值的动态范围;采用多特征判断法即通过灰度阈值、差分阈值、尺寸阈值3个方面判断检测到的目标是否为光斑。实验结果表明,该方法能有效增强图像对比度,改善图像质量,识别并滤除光斑,从而提高目标检测的有效性。

英文摘要：

For unclear underground monitoring video image and error target detection caused by low intensity of illumination,a lot of dust and strong light interference,an image preprocessing method for underground monitoring video was proposed.Histogram equalization algorithm is introduced to transform gray value from non-uniform distribution state to uniform distribution state and extend dynamic range of pixel gray value.Multi-features judgment method based on gray threshold,difference threshold and size threshold is used to identify spot.The experimental results show that the method can effectively enhance image contrast,improve image quality,and identify and filter spot,so as to improve effectiveness of target detection.