With the development of technical means of increasing popularity gained in situ of research that allow to observe the development of the process in time. This is particularly important for processes that occur locally. Corrosion of metals is characterized by the occurrence of local events on the surface - the formation of pits, gas bubbles, the development of cracks. Monitoring the progress of these external manifestations of corrosion leads to a deeper understanding of its mechanismand its dependence. Like most natural phenomena by corrosion it is substantially difficult and irregular, and therefore the morphology of the image surface corrosion (ISP) may not be completely identical, even when using the same material and corrosive environment. On the other hand, the ISP can be classified by the nature of Destruction - pitting, stress-corrosion, uniform, with gas evolution. The aim of this study was to develop an algorithm that will handle the ISP and quantitatively describe them. ISP created with the specified operator intervals during the experiment, and were maintained on a computer. Then each image is analyzed by a software developed for a specific type of corrosion. Imaging occurs in several stages: preparation of images, image binarization and edge detection of objects, quantitative characteristic of the obtained contours. The first treatment step in situ ICP is to prepare the image. At this stage, the removal of noise caused by the optical system and used for image capture equipment. In the second stage treatment is conducted on site ICP binarization image, i.e. the division of the image in the background (white) and information part (black). At the stage of allocation information of the image it is important to use morphological operations: building, erosion, opening and closing. In the third stage the researcher task allocation contours of objects in the ICP. To do this, use the language to describe images. For a description of the boundaries and islands used chain coding scheme suggested by Freeman. Thus, the image is represented by a chain (sequence) or chains octal numbers. Changing the grid granularity, superimposed on the image, you can get any desired resolution. In the fourth step of processing is carried out ICP analysis features found objects. Specific information is determined depending on the task. It can be a number of selected objects, their perimeter, area, density, fractal dimension. In the examples given specific image processing program: 1) an aluminum surface in 0.1 M NaCl (pH 11) at potential corrosion; 2) The steel surface in 0.1 M NaCl (pH 7) against borate buffer in the anode polarization conditions; 3) of the section cross-section of corrosion cracks formed on the actual main gas pipeline during its operation.
展开▼
机译：随着技术手段的发展，越来越多的现场研究获得了普及，可以及时观察过程的发展。这对于本地发生的过程尤其重要。金属腐蚀的特征是在表面上发生局部事件-形成凹坑，气泡，形成裂纹。监测腐蚀的这些外部表现的进程可以使人们对其腐蚀机理及其依赖性有更深入的了解。像大多数自然腐蚀现象一样，这非常困难且不规则，因此即使使用相同的材料和腐蚀性环境，图像表面腐蚀（ISP）的形态也可能不完全相同。另一方面，可以根据破坏的性质对ISP进行分类-点蚀，应力腐蚀，均匀，有气体逸出。这项研究的目的是开发一种可以处理ISP并定量描述它们的算法。 ISP在实验期间以指定的操作员间隔创建，并在计算机上维护。然后，通过针对特定腐蚀类型开发的软件对每个图像进行分析。成像分为几个阶段：准备图像，图像二值化和对象的边缘检测，所获得轮廓的定量特征。原位ICP的第一步是准备图像。在这一阶段，消除了由光学系统引起的并用于图像捕获设备的噪声。在第二阶段中，处理是在现场进行ICP二值化图像，即在背景（白色）和信息部分（黑色）中进行图像分割。在图像的分配信息阶段，重要的是使用形态学操作：构建，侵蚀，打开和关闭。在第三阶段，研究人员在ICP中分配对象的轮廓。为此，请使用语言描述图像。对于边界和岛的描述，使用了Freeman建议的链编码方案。因此，图像由一个链（序列）或一个链八进制数表示。更改叠加在图像上的网格粒度，可以获得任何所需的分辨率。在处理的第四步中，进行ICP分析功能找到的对象。具体信息取决于任务。它可以是许多选定的对象，它们的周长，面积，密度，分形维数。在给出特定图像处理程序的示例中：1）铝表面在0.1 M NaCl（pH 11）中处于潜在腐蚀状态; 2）在阳极极化条件下，用硼酸缓冲液在0.1 M NaCl（pH 7）中的钢表面; 3）截面图实际运行中在主燃气管道上形成的腐蚀裂纹的截面。
展开▼
