智能系统学报
智能繫統學報
지능계통학보
CAAI Transactions on Intelligent Systems
2015年
5期
803-809
,共7页
陈珂%柯文德%许波%张良均
陳珂%柯文德%許波%張良均
진가%가문덕%허파%장량균
图像去雾%去雾效果评价%大气散射模型%相对色彩空间%可见边对比度
圖像去霧%去霧效果評價%大氣散射模型%相對色綵空間%可見邊對比度
도상거무%거무효과평개%대기산사모형%상대색채공간%가견변대비도
image defogging%defogging effect assessment%atmospheric scattering model%opponent color space%contrast of visible edges
针对当前对图像去雾效果评价的不足,提出了一种改进的评价彩色图像去雾效果的方法. 该方法同时考虑了对图像边缘的评价以及对颜色失真的评价,基于图像雾化的大气散射模型,通过将原始图像转换到相对色彩空间,提出了度量颜色失真的标准;结合对比度增强的评价方式,提出了一个统一的评价指标,从而实现很好地给出一个符合人眼视觉判断的客观评价结果. 实验中基于多种去雾算法的去雾结果,对基于可见边比的评估方法、CNC评价指标和本文提出的评价指标进行了对比,结果表明本文改进的评价标准能更好地体现去雾的质量,获得与视觉判定更加接近的结论.
針對噹前對圖像去霧效果評價的不足,提齣瞭一種改進的評價綵色圖像去霧效果的方法. 該方法同時攷慮瞭對圖像邊緣的評價以及對顏色失真的評價,基于圖像霧化的大氣散射模型,通過將原始圖像轉換到相對色綵空間,提齣瞭度量顏色失真的標準;結閤對比度增彊的評價方式,提齣瞭一箇統一的評價指標,從而實現很好地給齣一箇符閤人眼視覺判斷的客觀評價結果. 實驗中基于多種去霧算法的去霧結果,對基于可見邊比的評估方法、CNC評價指標和本文提齣的評價指標進行瞭對比,結果錶明本文改進的評價標準能更好地體現去霧的質量,穫得與視覺判定更加接近的結論.
침대당전대도상거무효과평개적불족,제출료일충개진적평개채색도상거무효과적방법. 해방법동시고필료대도상변연적평개이급대안색실진적평개,기우도상무화적대기산사모형,통과장원시도상전환도상대색채공간,제출료도량안색실진적표준;결합대비도증강적평개방식,제출료일개통일적평개지표,종이실현흔호지급출일개부합인안시각판단적객관평개결과. 실험중기우다충거무산법적거무결과,대기우가견변비적평고방법、CNC평개지표화본문제출적평개지표진행료대비,결과표명본문개진적평개표준능경호지체현거무적질량,획득여시각판정경가접근적결론.
Currently, there is no adequate evaluation method for image defogging effect. Aiming at this, an im-proved method for the color image defogging effect assessment is proposed. By this method, either the edge' s con-trast enhancement or the color distortion is considered. Based on the atmospheric scattering model of an atomized image, the method first converts the original image to the opponent color space, thus deriving a standard for meas-uring the color distortion. Then, by combining the evaluation means for contrast enhancement, a unified assessment index for color image is generated, with objective evaluation results and good human visual perception. The defog-ging effects of many defogging methods were evaluated by comparison among the visible edge ratio, CNC value, and the criteria proposed in this paper. The results show that the improved assessment criteria can better reflect defog-ging quality and is closer to expected visual judgment.