CAJ | 학술논문

在浓雾环境下采集的视觉图像出现突兀点,需要进行平滑处理,提高对浓雾环境下的图像成像和识别能力.传统方法中采用差频控制的浓雾图像突兀点平滑处理算法,图像采集出现视觉偏移时平滑处理效果不好.提出一种基于视觉偏移修正的浓雾图像突兀点平滑处理算法,基于浓雾天气条件下单色大气散射模型,物体表面的反射光由于大气粒子的散射而损失,得到雾点散射损失,采用视觉偏移修正的算法,提高对浓雾图像的视觉偏移修正能力,由于起点坐标和视线方向出现视觉偏移,采用去雾滤波算法提高突兀点的平滑性能,得到浓雾图像的去雾滤波系统函数,通过边缘检测得到图像的突兀点平滑结构,进行Hough变换直线检测,实现视觉偏移修正和突兀点平滑改进.仿真结果表明,该算法具有较好的突兀点平滑处理性能,无论是细节还是浓雾图像的色彩都恢复得比较自然和清晰,去雾效果较好,平滑处理的图像有较高的配准率,展示了该算法的优越性能.
재농무배경하채집적시각도상출현돌올점,수요진행평활처리,제고대농무배경하적도상성상화식별능력.전통방법중채용차빈공제적농무도상돌올점평활처리산법,도상채집출현시각편이시평활처리효과불호.제출일충기우시각편이수정적농무도상돌올점평활처리산법,기우농무천기조건하단색대기산사모형,물체표면적반사광유우대기입자적산사이손실,득도무점산사손실,채용시각편이수정적산법,제고대농무도상적시각편이수정능력,유우기점좌표화시선방향출현시각편이,채용거무려파산법제고돌올점적평활성능,득도농무도상적거무려파계통함수,통과변연검측득도도상적돌올점평활결구,진행Hough변환직선검측,실현시각편이수정화돌올점평활개진.방진결과표명,해산법구유교호적돌올점평활처이성능,무론시세절환시농무도상적색채도회복득비교자연화청석,거무효과교호,평활처리적도상유교고적배준솔,전시료해산법적우월성능.
Visual image acquisition in fog environment emerges, the need for smoothing, to improve imaging and recognition ability of the fog environment. In the traditional method using the difference of the image frequency control abrupt point smoothing algorithm, image acquisition vision shifts when the smoothing effect is not good. A fog image visual offset correc-tion is proposed based on abrupt point smoothing algorithm, monochrome atmospheric scattering model based on the dense fog weather conditions, the surface reflected light due to scattering of atmospheric particles and loss, to get the fog point scattering loss correction algorithm using visual offset, improve visual offset correction ability of the image, because the starting point coordinates and gaze direction appear visually offset, used to fog filtering algorithm of smoothness abrupt point, get the image defogging filter system function, get the abrupt point smoothing image structure through the edge detec-tion, Hough transform line detection, realize the visual offset correction and abrupt point smoothing improvement. The simu-lation results show that the algorithm has better abrupt point smoothing processing performance, both the details or the fog color images to restore more natural and clear the fog, to good effect, image smoothing processing with high registration rate, show the superior performance of the algorithm.