Abstract Crack detection is an important task to monitor the structural health of the infrastructure exposed to fire. The manual inspections are time consuming and the outcome depends on the knowledge of the inspector. In order to overcome the subjective nature of the outcome, a method for detecting thermal cracks using Ripplet transform is proposed. The major components of the proposed method are noise removal, image enhancement, crack detection and detection of crack geometric parameters. Discrete Ripplet Transform (DRT) is used to identify the cracks in concrete subjected to elevated temperature. The bilateral filter is applied to the Ripplet coefficients to remove the noises in the image. Min–max gray level enhancement scheme is used to enhance the contours of the image. The crack pixels and background pixels are detected in the Ripplet domain in order to identify the major and minor thermal cracks completely. The geometric parameters of detected cracks are quantified through length, width, perimeter and area. The novelty of the proposed method depends on the usage of Ripplet transform for thermal crack detection. The results of the proposed thermal crack detection are compared with the other transform domain based state-of-the-art methods.

中文翻译：

---

使用涟漪变换评估火灾暴露混凝土结构的热裂纹

摘要 裂纹检测是监测暴露于火灾的基础设施结构健康状况的一项重要任务。人工检查非常耗时，结果取决于检查员的知识。为了克服结果的主观性，提出了一种利用波纹变换检测热裂纹的方法。该方法的主要组成部分是噪声去除、图像增强、裂纹检测和裂纹几何参数检测。离散波纹变换 (DRT) 用于识别受高温影响的混凝土中的裂缝。双边滤波器应用于 Ripple 系数以去除图像中的噪声。最小-最大灰度增强方案用于增强图像的轮廓。在波纹域中检测裂纹像素和背景像素，以完全识别主要和次要的热裂纹。检测到的裂缝的几何参数通过长度、宽度、周长和面积进行量化。所提出方法的新颖性取决于使用波纹变换进行热裂纹检测。将所提出的热裂纹检测的结果与其他基于变换域的最新方法进行了比较。