Analysis of Leaks in Flood Embankments Using Deterministic Methods and Computational Intelligence Algorithms

This article presents the concept of an information system based on electrical impedance tomography for monitoring flood embankments. To design such a solution, knowledge and experience from many fields is required. One example is the use of electrical tomography for non-invasive observation of objects such as reservoir tanks. An important part of the described system is the IoT platform, which includes devices with interfaces for exchanging current data and for collecting and storing historical data. The IoT platform provides access to data through analytical systems, providing client applications for processing information collected by the system. The article presents three image reconstruction algorithms, such as Gauss-Newton, level set function and multiply artificial neural network, whose main feature is the use of a uniform vector of input signals for separate training of individual neural networks, each of which generates a single pixel of the output image. The effectiveness of the hybrid method of input variables reduction combining ANN and EIasticNET was also examined. The results of the presented research confirm the high efficiency of the described solutions.