基于暗原色先验的高空定点图像去雾算法

High Altitude Fixed Point Images Defogging Based on Dark Channel Prior

室外高空定点监控系统、卫星遥感系统、无人机航拍系统等以其实时获取图像数据的优势,在工业中获得越来越广泛的应用。然而在雾霾气候下,这类系统拍摄的图像,由于大气光的散射作用,图像整体呈灰白色,清晰度降低,物体特征被掩盖难以辨认,影响图像后期的处理。针对这一问题,总结高空定点图像的主要特点,改进一种基于暗原色先验的高空定点图像去雾算法,该算法通过分割思想,将图像分为天空与非天空部分,并对这两个部分采用不同的去雾方案。实验结果显示,相对于原算法,改进算法在天空区域平滑过渡,整体去雾效果清晰自然,去雾图像经客观质量评价较原算法有明显提升。

Outdoor high altitude fixed point monitoring system, satellite remote sensing system, UAV aerial system with its real-time access to image data advantages in the industry to obtain more and more widely used. However, in the haze climate under the image of such systems, the scattering of atmospheric light to reduce the image quality, the overall image was gray, reduced clarity, object features are difficult to identify the cover, affecting the image of the latter part of the treatment. Aiming at this problem, summarizes the main characteristics of high altitude fixed point images, and improves a high altitude fixed point image defogging algorithm based on dark primary color a priori. The algorithm divides the image into sky and non-sky parts by dividing the idea. The two parts use different defogging schemes. The experimental results show that compared with the original algorithm, the improved algorithm has a smooth transition in the sky area, and the overall fog effect is clear and natural. The fog image is improved by the objective method.