针对镁熔液第一气泡随机析出图像的检测问题,借助信息熵、纹理、局部信杂比等复杂度因子来定量地描述第一气泡在不同位置析出时邻域背景复杂度和全域背景复杂程度。通过分析背景复杂度高低,设计通用的图像检测方法,研究镁熔液第一气泡随机析出图像检测方法的普遍适用性。实验结果表明:第一气泡在第一位置析出时邻域背景复杂度和全域背景复杂度均最大,且背景复杂度高的图像检测方法适用于其他背景复杂度低的图像,为镁熔液第一气泡随机析出图像处理提供了一种新思路。
For the detecting problem of random precipitation images of the first bubble in magnesium alloy melt, complexity factors including information entropy, texture, local signal - to - clutter ratio, and other relevant factors are utilized to quantitatively describe neighborhood background complexity and global background complexity when the first bubble precipitates from different locations. General image processing algorithms are designed, and the uni- versal applicability of image processing algorithms about random precipitation of the first bubble. Experiment results show that neighborhood background complexity and global background complexity of the bubble when precipitating from the first location are the highest and the image processing algorithms of high background complexity are applicable to the images of other low background complexity, providing a novel way about image processing of random pre- cipitation of the first bubble in magnesium alloy melt.