Vestnik of Samara State Technical University. Technical Sciences Series

The article focuses on the development of a decision support system based on computer vision technologies and image preprocessing methods. It presents a comparative analysis of algorithms used to enhance the quality of visual data and to automatically identify key objects in images. The study emphasizes various methods such as filtering, equalization, and image segmentation, which are designed to improve the accuracy of computer vision algorithms in detecting important structures. The research tested and evaluated several image preprocessing techniques, including adaptive and standard histogram equalization, median filtering, and gamma correction. The results demonstrate that applying image preprocessing significantly improves the quality of data analysis. The Shi-Tomasi algorithm showed the highest efficiency in object recognition, especially when used with equalization techniques, allowing for precise identification of structural landmarks. The paper highlights the critical role of preprocessing in boosting the performance of computer vision systems across various domains, such as industrial applications, quality control, safety, and autonomous systems. These systems enable the automation of image analysis while reducing the impact of human error in decision-making processes. A practical example of these technologies is the development of a system for diagnosing musculoskeletal injuries and pathologies based on X-ray images. The implementation of such systems in medical practice accelerates diagnostic processes and improves the accuracy of diagnoses, ultimately reducing the workload of specialists and enhancing the quality of healthcare services.

The article discusses a method for monitoring the condition of electric drive equipment, which will allow you to monitor its operating mode in real time. During operation, it is proposed to take into account the operational characteristics of the electric drive. As part of the work, the design of the conveyor belt was analyzed in terms of its reliability. Based on the known dependencies, mathematical models have been developed that describe the operating time to failure of the elements, which helps to increase the adequacy of the assessment of the service life of the equipment when evaluating the factors of its operation. To carry out the research, a block diagram of the automatic control system for diagnosis and restoration of the electric drive was synthesized. This approach demonstrates progressive and modern ways to ensure the reliability of technological equipment, which, in turn, optimizes the operation of the maintenance and repair system based on the actual condition of the equipment. The proposed diagnostic system plays an important role in reducing downtime and reducing operating costs when using electric drives in enterprises. With its help, enterprises will be able to plan maintenance more efficiently and provide electric drive resources in the most efficient way, which will lead to an increase in overall productivity and reduce the risks associated with sudden failures.

The relationship between the input and output signals of a chromatograph is demonstrated. The expediency of improving the signal-to-noise ratio and using correlation data processing for this purpose is noted. An algorithm for calculating a chromatogram averaged over multiple injections of a sample from a specific individual specimen is presented. This algorithm is then modified for the case of calculating a differential averaged chromatogram, which is required to determine the difference in the component composition of two different specimens. A Legendre pseudorandom sequence is used to generate the input signals. The alphabet of this sequence is 1 or 0: 1 indicates a sample injection, and 0 indicates the absence of a sample injection at a given time. In the case of analyzing two specimens, the input sequences controlling the injection of the first and second specimens are complementary: if the sample of the first specimen is not injected, the sample of the second specimen is injected, and vice versa. The algorithm for calculating the averaged chromatogram is reduced to regression analysis. An information matrix of the experimental design is introduced. The accuracy of estimation is characterized using the covariance matrix of estimates. The choice of the optimal input sequence for sample injection is based on the D-optimality condition (the condition for maximizing the determinant of the information matrix), which ensures the minimum generalized variance of all estimates of the ordinates of the averaged chromatogram. The results of a numerical simulation experiment are presented, proving the effectiveness of using the correlation data processing method for averaging chromatograms.

This paper proposes integrating an energy gas turbine expander unit (EGTEU) and a proton exchange membrane (PEM) electrolysis unit into the standard technological scheme of a gas distribution station (GDS) for main gas pipelines (MGP). This solution represents an effective systematic approach to modernizing standard GDS, which not only increases their energy efficiency but also significantly reduces their environmental impact. A thermal scheme has been developed, and the operating principle of the new type of GDS is described. Based on thermodynamic analysis, the potential for generating additional electric power in the EGTEU to produce hydrogen in the PEM electrolysis unit has been evaluated.

By modernizing standard GDS, it is possible to achieve electric power generation with an efficiency of up to 60 %, which is sufficient to power ultra-large-scale hydrogen production facilities. This presents promising opportunities for producing methane-hydrogen mixtures at GDS as a prospective fuel for consumers, which can significantly improve environmental performance compared to traditional natural gas.