CAJ | 학술논문

无人机在进行情报侦察时，容易受到不良天候条件的影响，使得侦察图像模糊不清，对比度下降，但是现有较多算法解决此类问题的实时性较差，所以本文提出了基于暗通道原理的无人机侦察图像快速去雾算法。首先对图像进行降采样，根据大气物理模型和暗通道原理，利用最小值滤波和均值滤波分析出透射率和大气光强的求解方法并进行优化，最后求解出去雾图像。通过实验验证，本文算法简单有效，实时性极高，可以用于工程实现。而且算法对图像的亮度也有一定的调整功能，具有较强的鲁棒性。
무인궤재진행정보정찰시，용역수도불량천후조건적영향，사득정찰도상모호불청，대비도하강，단시현유교다산법해결차류문제적실시성교차，소이본문제출료기우암통도원리적무인궤정찰도상쾌속거무산법。수선대도상진행강채양，근거대기물리모형화암통도원리，이용최소치려파화균치려파분석출투사솔화대기광강적구해방법병진행우화，최후구해출거무도상。통과실험험증，본문산법간단유효，실시성겁고，가이용우공정실현。이차산법대도상적량도야유일정적조정공능，구유교강적로봉성。
When the UAV is processing the intelligence and surveillance, it is easily affected by the bad weather, leading to the decrease of image definition and contrast. This paper raises the fast haze removal algorithm based on the dark channel of UAV reconnaissance image. Firstly, down-sampling the image and according to the physical model of atmosphere and the dark channel theory, the algorithm analyzes the method that how to do with the transmissivity and the intensity of the atmospheric light by the mini-value filtering and average filtering, and then optimizes the method. At last it receives the de-hazing image. It is proved that the algorithm is simple and effective and has the high real-time performance by a lot of experiments. It can be used for the project realization. And to a certain extent it is able to adjust the brightness of the image and is with strong robustness.