摘要
在复杂的石化爆燃环境中,巡检机器人进行指针式仪表检测时,相机容易发生相对运动造成图像产生运动模糊,导致仪表等识别精度降低。为解决此类问题,提出一种改进的模糊图像点扩散函数(PSF)的模糊参数估计方法。对图像进行预处理后,采用Canny算法对图像进行边缘检测,筛选出边缘信息较多的子图像块;之后对图像进行倒频谱分析,采用Canny算法对倒频谱图进行边缘检测,通过Radon变换求解模糊角度,根据倒频谱负峰值距离估计出模糊尺度;基于估计的PSF参数,采用维纳滤波完成图像复原;为降低指针提取的难度,对复原图像进行自适应阈值分割与图像细化,利用Hough变换检测出仪表表盘及指针,完成自动读数。
In complex petrochemical deflagration environment, when the inspection robot performs pointer-type instrument detection, the camera is prone to occur relative motion causing the image to produce motion blur, which reduces the recognition precision of instrument.To solve such problems, an improved blur parameter estimation method of the point spread function(PSF)of blurred images is proposed.After preprocessing the image, the Canny algorithm is used to detect the edges of the image, and filter out the sub-image blocks with more edge information.Then perform cepstrum analysis on the image, the Canny algorithm is used to detect the edges of the cepstrum, Radon transform is used to solve the blur angle, and estimate the blur length based on the distance of the negative peak of the cepstrum, based on the estimated PSF parameters, Wiener filtering is used to complete image restoration.In order to reduce the difficulty of pointer extraction, adaptive threshold segmentation and image refinement are performed on the restored image, and the instrument dial and pointer are detected by Hough transform to realize automatic reading.
作者
孙绍祖
刘今越
史宝军
SUN Shaozu;LIU Jinyue;SHI Baojun(School of Mechanical Engineering,Hebei University of Technology,Tianjin 300401,China)
出处
《传感器与微系统》
CSCD
北大核心
2022年第9期109-112,120,共5页
Transducer and Microsystem Technologies
基金
国家重点研发计划资助项目(2019YFB1312100)
河北省重点研发计划资助项目(20311803D)
河北省自然科学基金资助项目(E2019202338)。
