Edge detection for ultrasound-based, non-destructive examination of nuclear power plants

管道完整性检查对炼油厂、矿山和核电站的维护至关重要。管道检测通常依赖于能够实现管道三维可视化的技术，以及随后对管道尺寸、形状和厚度的测量。安大略发电公司（OPG）开发了一种基于超声波的成像技术，用于评估核反应堆馈电管道的完整性。从超声图像中测量管道完整性的关键步骤是管道边界的检测。在这个项目中，我们将利用我们在定量图像分析方面的专业知识，开发新的方法，从给料管道的超声图像中描绘管道内外边界。我们将探索基于三种不同算法的解决方案：自适应阈值，流传播活动轮廓和分水岭变换。新方法将使用模拟和现场数据进行评估，我们将定量比较我们的方法

The inspection of pipeline integrity is critical to maintain refineries, mines and nuclear plants. Pipeline inspection often depends on technologies that allow visualization of the pipe in three dimensions, and subsequent measurement of size, shape, and thickness. Ontario Power Generation (OPG) has developed an ultrasound-based imaging technology to evaluate the integrity of feeder pipes in nuclear reactors. A critical step in the measurement of pipe integrity from ultrasound images is the detection of pipe boundaries. In this project, we will leverage our expertise in quantitative image analysis to develop new methods to delineate the internal and external pipe boundaries from ultrasound images of feeder pipers. We will explore solutions based on three different algorithms: adaptive thresholding, flow-propagated active contours and watershed transformations. The new methods will be evaluated using simulated and field data, and we will quantitatively compare our