夜間增強技術應用於一般閉路電視監測影像之研究

Abstract

現今夜間安全監控的議題越來越重要，有心人士企圖利用夜間燈光不佳，監視系統無法完整發揮作用的情況下，進行某些犯罪行為，造成民眾財物損失及心理的負擔，另外在一些公共區域也經常架設攝影機進行監控，並依監視器畫面進行後端決策。針對以上情況，一般攝影機在夜間環境所捕捉的畫面相較不太清楚，無法提供良好的資訊。本論文針對上述問題，設計一套適用於夜間監控有所幫助的系統，主要目的在於增加夜間場景的清晰度，免除夜間影像無法進行識別的死角，減少觀測人員的視覺疲勞，並提高監控系統的應用性。 本論文的貢獻有以下三點：第一，本論文提出一種影像混合方法，主要包括白天背景照明影像與夜間照明增強影像，使得周遭環境和目標物於黑暗環境下更能清楚識別; 第二，經由前景資訊設置權重，將照明影像進行混合，達到夜間燈光保留及亮度均勻之成效。；第三，設計一個夜間增強的監控系統，在平均亮度大於30的限制條件下，於若干公共環境中證實本系統能達到良好的效果。

The issue of nighttime surveillance become more and more important. When surveillance system doesn't function well at night, some people take advantage the occasion to steal or destroy others' belonging. Besides, cameras are often used to set up public to let the professionals making right decisions. In view of these situations, the design of nighttime surveillance would be helpful to provide clear images, reduce visual fatigue for observer. Moreover, to bring the surveillance system into practical. There are three contributions of this thesis. First, we propose a image fusion method, including daytime and nighttime background illumination to make the targets and surrounding more visible in poor light. Second, we set weight function by foreground and get uniform illumination and by fuse image. Finally, we complete the prototype of system in enhance nighttime content. In constraints which brightness over 30 prove our system can work well in several public area.