压水堆燃料辐照后燃料棒棒间距数据机器视觉测量方法研究

针对辐照后燃料棒棒间距数据获取和处理困难的问题，基于燃料棒几何特性及其在压水堆燃料组件中的排列方式，本文提出一种基于机器视觉的高效、可靠的燃料棒棒间距数据测量方法。该方法首先采用Retinex算法对水下燃料棒的采集图像进行增强预处理；然后，针对燃料棒阵列的前后成像干扰问题，采取边缘增强和逐行灰度特征处理方法实现待测燃料棒与背景燃料棒的有效分离，并进一步提升图像清晰度；最后，对燃料棒图像的单行灰度值进行二次曲线拟合，获得各个燃料棒的亚像素边缘点坐标。乏燃料组件的现场实验验证结果表明，该方法可一次性实现16个燃料棒棒间距测量，且测量精度达±0.32 mm，可为燃料性能分析提供高效、可靠的数据支持。 

Research on Machine Vision Measurement Method of Rod Spacing Data for PWR Fuel after Irradiation

To address the difficulty of obtaining and processing fuel rod spacing data after irradiation, based on the geometric characteristics of fuel rods and their arrangement in PWR fuel assemblies, this paper proposes an efficient and reliable method for measuring fuel rod spacing data based on machine vision. The method first uses Retinex algorithm to enhance the preprocessing of the collected images of underwater fuel rods; then, for the front and rear imaging interference problem of the fuel rod array, edge enhancement and line-by-line grayscale feature processing methods are used to effectively separate the fuel rods to be tested from the background fuel rods, and further improve the image clarity; finally, the single-row gray value of the fuel rod image is fitted with quadratic curve to obtain the sub-pixel edge point coordinates of each fuel rod. The field experimental verification results of spent fuel assemblies show that this method can realize the rod spacing measurement of 16 fuel rods at one time, and the measurement accuracy is up to ±0.32 mm, which can provide efficient and reliable data support for fuel performance analysis.