CAJ | 학술논문

针对目前乙烯气体泄漏检测存在效率低、精度差以及无法准确标定泄漏位置等问题，提出了一种基于红外成像多特征联合的乙烯气体泄漏检测方法。该方法首先利用红外相机获取生产车间的红外视频图像，通过联合帧间差分和背景差分的方法获取疑似区域；其次对疑似区域进行泄漏气体成像的统计特征和小波能量变化特征分析，找出特征阈值；最后将多个特征阈值条件联合在一起，作出是否存在泄漏的判定。实验结果表明：该方法准确、高效且具备实时性，其泄漏检测的准确率达到95.6%。
침대목전을희기체설루검측존재효솔저、정도차이급무법준학표정설루위치등문제，제출료일충기우홍외성상다특정연합적을희기체설루검측방법。해방법수선이용홍외상궤획취생산차간적홍외시빈도상，통과연합정간차분화배경차분적방법획취의사구역；기차대의사구역진행설루기체성상적통계특정화소파능량변화특정분석，조출특정역치；최후장다개특정역치조건연합재일기，작출시부존재설루적판정。실험결과표명：해방법준학、고효차구비실시성，기설루검측적준학솔체도95.6%。
Aiming at the problems of low efficiency,precision,and that the leak location cannot be accurately calibrated in ethylene gas leakage detection,present a method for detecting ethylene gas leakage based on multiple features combination of infrared imaging. Firstly,Infrared image of the production workshop is obtained by the infrared camera,and the suspected area is obtained by the method of the combination of frame difference and background difference;secondly,the statistics features and the wavelet energy variation characteristics of gas leakage imaging in suspected area is analyzed to find out the feature threshold;finally,the multiple feature threshold conditions is combined together,to determine whether there is leakage. The experimental results show that this method is accurate,efficient and real-time,and the leak detection accuracy rate reaches 95. 6%.