地下水动力场演化图像奇异点检测方法研究  

Research on Singularity Detection Method in Groundwater Dynamic Field Evolution Images

作　　者：李默寒 LI Mohan(Ecological Environment Geology Service Center,Henan Province Geology Bureau,Zhengzhou 450000,China)

机构地区：[1]河南省地质局生态环境地质服务中心,河南郑州450000

出　　处：《自动化仪表》2024年第3期108-112,共5页Process Automation Instrumentation

摘　　要：为了增强图像中的细节、完善平滑处理、提高图像中奇异点的检测精度,提出基于改进Renyi熵的地下水动力场演化图像奇异点检测方法。通过求解地下水动力场演化图像的梯度场,计算不同亮度下图像的相对梯度。利用图像增强函数,对地下水动力场演化图像作增强处理。创新性地根据三维直方图将地下水动力场演化图像划分为图像背景域和目标域。利用改进Renyi熵,定义了演化图像在目标域和背景域的三维Renyi熵。结合最大熵分割原理,分割了地下水动力场演化图像。引入平滑尺度,对地下水动力场演化图像进行平滑处理。利用概率函数,建立图像奇异点检测函数。设计了地下水动力场演化图像奇异点检测算法。试验结果表明,所提方法能够准确检测图像中的奇异点,并将检测熵值和检出率分别提高到0.8以上和90%以上。该研究有效提高了地下水动力场演化图像中奇异点的检测性能。To enhance the details in the image,perfect the smoothing process,and improve the detection accuracy of singularities in the image,the singularity detection method in groundwater dynamic field evolution image based on improved Renyi entropy is proposed.By solving the gradient field of the groundwater dynamic field evolution image,the relative gradients of the image under different brightness are calculated.The image enhancement function is utilized to enhance the processing of the groundwater dynamic field evolution image.Innovatively,the groundwater dynamic field evolution image is divided into image background domain and target domain according to the three-dimensional histogram.The three-dimensional Renyi entropy of the evolution image in the target and background domains is defined by using the improved Renyi entropy.Combined with the principle of maximum entropy segmentation,the groundwater dynamic field evolution images are segmentd.The smoothing scale is introduced to smooth the evdution image of groundwater dynamic field.The probability function is used to establish the image singularity detection function.The singularity detection algorithm of the evolution image of groundwater dynamic field is designed.The experimental results show that the proposed method can accurately detect the singularities in the image and improve the detection entropy value and detection rate to more than 0.8 and more than 90%,respectively.This study effectively improves the detection performance of singularities in groundwater dynamic field evolution images.

关 键 词：奇异点检测 地下水 演化图像 改进Renyi熵 动力场 图像增强 梯度场 

分 类 号：TH-39[机械工程]