小波域BM3D滤波算法高炉回旋区温度场测量

Measurement of temperature field in blast furnace raceway based on BM3D filtering algorithm in wavelet domain

高炉风口回旋区温度的精准检测对判断和控制高炉运行状况至关重要。为实现温度场的数字化测量,避免人工观测的主观性,该文提出一种小波域BM3D辐射图像滤波算法。针对风口图像中条纹噪声与高斯噪声的影响,首先利用小波分解实现有用信息与噪声分离;其次将携带条纹噪声与高斯噪声的细节信息分别采用软阈值与BM3D进行滤波。这样分步式的滤波算法在保证高斯噪声的去噪效果的同时,可有效抑制条纹噪声。去噪测试结果为MSE为10.1158,PSNR为38.0808 dB。测温实验结果表明,该文提出的算法测得的温度场准确度相对更高。

Accurate measurement of the temperature in the tuyere and raceway of blast furnace is very important to judge and control the running condition of blast furnace.In order to realize digital measurement of temperature field and avoid subjectivity of manual observation,a combined filtering algorithm based on wavelet transform and BM3D algorithm is proposed in this paper.Aiming at the influence of fringe noise and Gaussian noise in the image of tuyere,this paper firstly uses wavelet decomposition to separate useful information from noise.Secondly,the detail information of fringe noise and Gaussian noise was filtered by soft threshold and BM3D respectively.This step by step filtering algorithm can not only ensure the effect of Gaussian noise removal,but also effectively suppress the fringe noise.The MSE and PSNR were 10.1158 and 38.0808 dB respectively.The experimental results show that the algorithm proposed in this paper has higher accuracy in measuring the temperature field.