基于智能图像的大气污染灾害等级判定研究

针对当前大气污染灾害等级评价方法存在准确性和实时性差等问题,提出了基于智能图像的大气污染灾害等级鉴定方法。给出含噪大气污染图像,依据图像自身信号强度与噪声信号强度的差异性,利用噪声极值检测出图像中的噪声点。基于图像受到的噪声污染程度自适应选取不同尺寸图像滤波窗口,利用自适应中值滤波方式增强大气污染图像。将增强后的图像引入至大气污染灾害等级鉴定中,分别计算标准差、平均梯度、熵以及空间频率等能够反映大气污染图像变化的参数,并结合图像特征相似度衡量标准Canberra距离定义,构建大气污染灾害等级鉴定模型。将待鉴定大气污染图像代入模型中,并融合我国大气污染灾害等级判定标准,输出最终污染等级鉴定结果。实验结果表明,该方法污染等级鉴定准确率高,且具有较强地实时性。

Air Pollution Disaster Level Determination Based on Intelligent Image

Aiming at the problems of current air pollution hazard level evaluation methods,such as accuracy and poor real-time performance,we propose an intelligent image-based air pollution hazard level identification method.Based on image of noisy atmospheric pollution given and according to the difference between the signal strength of the image and the intensity of the noise signal,we detect the noise point in the image by the noise extreme value.The image filtering window of different sizes is adaptively selected based on the degree of noise pollution received by the image,and the atmospheric pollution image is enhanced by the adaptive median filtering method.The enhanced image is introduced into the air pollution hazard level identification,and the parameters such as standard deviation,average gradient,entropy and spatial frequency,which can reflect the change of atmospheric pollution image,are calculated respectively,and the image is constructed according to the Canberra distance definition of the image feature similarity measure.The image of the air pollution to be identified is substituted into the model,and the air pollution hazard level judgment standard is incorporated,and the final pollution level identification result is output.The experimental results show that the proposed method has high accuracy of pollution level i-dentification and strong real-time performance.The overall performance of the method is superior to the current related research and can provide reliable support for air pollution assessment.