摘要
针对雾霾天气下的光源运动目标清晰度低、轮廓模糊的问题,以火炮射击空中光源目标测量火炮跳角为研究背景,提出基于边窗均值滤波的雾霾天气下目标检测算法。首先在雾天退化模型基础上,使用边窗均值滤波算法对初始透射率进行改进,保留图像的轮廓边缘,然后通过平均修补块方法求出大气光值并复原出原始图像。最后通过奇异值分解的方法,判断光源运动目标和天空背景的相近程度,动态改变高斯混合模型的判别阈值。通过不同去雾算法的主观和客观分析以及光源微小运动目标的检测分析,所设计算法的目标检测清晰度更好。实验结果表明,该算法耗时较少,平均检测精度可达90%,能够准确、高效地检测出光源运动目标。
In order to solve the problems of low definition and fuzzy contour of moving target in haze weather,a target detection algorithm based on edge window mean filter is proposed in this paper.Firstly,the edge window means filtering algorithm based on the fog degradation model is used to improve the initial transmittance and preserve the contour edge of the image.Then,the atmospheric light value is calculated by the average patch and the original image is restored.Finally,singular value decomposition is used to judge the similarity between moving object and sky background,and the threshold of GMM is changed dynamically.Through the subjective and objective analysis of different defogging algorithms and the detection analysis of light source micro moving target,the target detection clarity of the designed algorithm is better.The experimental results show that the algorithm consumes less time,and the average detection accuracy can reach 90%.It can detect the moving object of light source accurately and efficiently.
作者
谢艳丽
姜志
王军
吴云鹏
XIE Yanli;JIANG Zhi;WANG Jun;WU Yunpeng(Suzhou University of Science and Technology,Suzhou Jiangsu 215009,China;Changchun Institute of Optics,Fine Mechanics and Physics,Chinese Academy of Sciences,Changchun 130033,China;Center of Arms Experiment of Baicheng,Baicheng Jilin 137001,China;Science and Technology on Near-Surface Detection Laboratory,Wuxi Jiangsu 214035,China)
出处
《激光杂志》
CAS
北大核心
2021年第11期46-52,共7页
Laser Journal
基金
“十三五”江苏省重点学科项目(No.20168765)、江苏省研究生科研创新项目(No.KYCX17_2060)、江苏省研究生工作站项目、近地面探测技术重点实验室基金(No.TCGZ2018A005)。
