一种适应光照多变的夜间前景提取方法
    点此下载全文
引用本文:叶慎飞,汪志成 ,张志君.一种适应光照多变的夜间前景提取方法[J].计算技术与自动化,2021,(1):129-134
摘要点击次数: 490
全文下载次数: 0
作者单位
叶慎飞,汪志成 ,张志君 (东华理工大学江西 南昌 330013) 
中文摘要:夜间前景提取是智能监控的必要处理过程,也是新时代下店铺防盗的高效手段。提出了一种在强弱光多变下具有检测鲁棒性的动态目标自动提取算法。在亮度通道上利用直方图均衡技术根据设置参数自适应灰度增幅,结合Canny边缘图像进行差分处理,最后给出前景提取算法。实验结果表明:在强弱光不同光照环境下,该方法可以很好的削弱不同光照对前景提取带来的多噪声、过度曝光等影响,提高前景对比度,相比其他方法,显示效果清晰完整,并满足实时性要求。
中文关键词:前景提取  直方图均衡  边缘图像  差分处理
 
A Night Foreground Extraction Method Suitable for Changing Light
Abstract:Nighttime foreground extraction is a necessary process for intelligent monitoring and an efficient means of anti-theft for shops in the new era. A dynamic target automatic extraction algorithm with robust detection under strong and weak light is proposed. On the brightness channel, the histogram equalization technique is used to adaptively increase the gray scale according to the set parameters, and the differential processing is performed in conjunction with the Canny edge image. Finally, the foreground extraction algorithm is given. The experimental results show that: under different light environments with strong and weak light, this method can well reduce the effects of different noise on the foreground extraction, such as multiple noise and overexposure, and improve the foreground contrast. Compared with other methods, the display effect is clear and complete. And meet the real-time requirements.
keywords:foreground extraction  histogram equalization  edge image  difference processing
查看全文   查看/发表评论   下载pdf阅读器