基于改进双边滤波的气体泄漏检测

针对传统泄漏检测方法效率低、鲁棒性差的问题，提出了一种基于红外图像技术的气体泄漏检测方法。该方法根据被测容器充气前后的红外图像的信息差异，对获得的热像图进行背景估计、差分，以及目标与噪声的辨别，利用基于改进的双边滤波算法实现对泄漏点的检测与定位。实验证明：双边滤波算法对背景的预测具有鲁棒性，通过目标与噪声的扩散性区别分辨出目标，实现对泄漏点的准确定位。

Air Leak Detection Based on Bilateral Filter of Infrared Image

 A gas leak detection method based on infrared image technology is proposed for problems of low efficiency and poor robustness of traditional leak detection methods. According to information difference of infrared images before and after a tested container is inflated, the method carries out background estimation, difference and target and noise discrimination on the obtained thermal image. Thus detection and location of a leakage point are realized by an improved bilateral filtering algorithm. The robustness on background prediction for the bilateral filtering algorithm is shown by the experimental results. The target is distinguished by the difference between diffusion of the target and the noise, thus realizing the accurate location of leakage point.