
The paper deals with the development of a model for real-time recognition and classification of UAVs and birds based on the training of the YOLOv10 neural network. The research area is considered relevant in connection with the problems of UAV detection in the context of security, given their growing use in various fields. A dataset consisting of 6,255 images collected from proprietary archives and public resources is trained to train the model. The process of data annotation, augmentation and distribution was implemented using Roboflow.com service. The model was trained on NVIDIA GeForce RTX 4080 GPU using Ultralytics framework. Test results showed high recognition accuracy with mAP50 and mAP50-95 metrics exceeding previous versions of YOLO. The model demonstrates the ability for efficient object segmentation and tracking, which makes it promising for optoelectronic surveillance applications. The results of the study can be useful for developers of UAV and bird detection and classification systems, as well as for improving safety in various fields. © 2025, Kazakh-British Technical University. All rights reserved.

This article explores the challenges of integrating two deep learning neural networks, YOLOv5 and RT-DETR, to enhance the recognition of unmanned aerial vehicles (UAVs) within the optical-electronic channels of Sensor Fusion systems. The authors conducted an experimental study to test YOLOv5 and Faster RT-DETR in order to identify the average accuracy of UAV recognition. A dataset in the form of images of two classes of objects, UAVs, and birds, was prepared in advance. The total number of images, including augmentation, amounted to 6337. The authors implemented training, verification, and testing of the neural networks exploiting PyCharm 2024 IDE. Inference testing was conducted using six videos with UAV flights. On all test videos, RT-DETR-R50 was more accurate by an average of 18.7% in terms of average classification accuracy (Pc). In terms of operating speed, YOLOv5 was 3.4 ms more efficient. It has been established that the use of RT-DETR as the only module for UAV classification in optical-electronic detection channels is not effective due to the large volumes of calculations, which is due to the relatively large number of parameters. Based on the obtained results, an algorithm for combining two neural networks is proposed, which allows for increasing the accuracy of UAV and bird classification without significant losses in speed.

This article investigates fraud detection in financial transaction networks using machine learning and graph-based typologies. The object of the study is financial transaction data, analyzed to improve the accuracy and efficiency of identifying fraudulent activities. The problem addressed is the limited generalizability and low recall of traditional fraud detection models in complex, real-world settings. To address this, a hybrid framework was developed that integrates Random Forests, neural networks, and graph-based typology indicators. Seven laundering typologies were extracted from a transaction graph – fan-in, fan-out, scatter-gather, gather-scatter, cycle, bipartite, and stacked bipartite – and used as additional features for classification. SMOTE was applied to correct class imbalance during training. Experimental results show that adding typology features significantly improves model performance. The best results were obtained with Random Forest: 98.5% accuracy, 79.1% precision, 56.3% recall, and an F1-score of 65.7%. Adding typology-based flags raised recall by 9–11 percentage points compared to models without them. Graph patterns like fan-in and fan-out were detected in 3.5–5.1% of transactions, while more complex structures such as cycle and scatter-gather appeared less frequently but correlated more strongly with known fraud. Unsupervised methods also showed promise: an autoencoder captured 60% of fraud cases among the top 2% anomalous transactions, while K-means identified 55% of fraud within flagged clusters. These methods proved useful for identifying emerging fraud types not yet labeled in training data. The model is suitable for integration into financial security systems with minimal input requirements – account IDs, timestamps, and transaction amounts – alongside basic graph analytics. Its robustness across datasets suggests strong applicability across diverse financial institutions. Copyright © 2025, Authors.

This work describes an automated system for preprocessing gravimetric data from an oil and gas deposit located in the southeastern part of the Caspian Depression. The functional modelling of the company’s gravimetric monitoring was accomplished. The modelling allowed us to define the functionality and models of the system data. The executive part and database of the system are compiled by using Oracle APEX. A specialized geographic information system (GIS) was developed as part of the visualization subsystem. The GIS uses Google Maps API as a basis for the visualization of layers and as a provider of maps

The results of electrical exploration carried out using the method of magnetotelluric sounding at a deposit in Western Kazakhstan are presented. Based on processing and interpretation of the electrometric data, a digital geoelectric model of the exploration area is developed and a geoelectric section of the test profile is constructed. It is shown that MT soundings, along with a seismic survey and well-log data, can be effectively used to define a geological structure and to identify prospective oil-and-gas bearing segments in Mesozoic terrigeneous-sedimentary complexes

There was shown that LLP "SPC "Geoken" along with best models of foreign electro-prospecting equipment from famous manufacturers (Zonge (USA), GDD and Phoenix Geophysics (Canada)) also successfully uses electro-prospecting equipment developed in Kazakhstan. There is some examples from real production projects completed by using this equipment. © 2018European Association of Geoscientists and Engineers EAGE

In this article, the authors review the practical experience of applying the VIP (vector induced polarization) method for geophysical surveying of one of the properties in Central Kazakhstan in 2017. The survey results demonstrate a significant reduction in time and cost associated with performing large area surveys achieved by using special IP arrays. The surveys carried out made it possible to rapidly detect abroad area of chargeability response which was followed up by deeper penetrating DDIP surveys to test for depth extent. © 2018European Association of Geoscientists and Engineers EAGE

Purpose. The research is aimed at creating a highquality geophysical basis for the additional geological study of Central Kazakhstan in determining the geological structure of the study area, identifying promising areas for further geological exploration, as well as analyzing the anomalous distribution of various minerals. Methodology. The research uses the methods of aeromagnetic, airborne gammaray spectrometry and gravimetric (ground) surveys. The obtained new airborne geophysical data are used for additional study of the geological structure of the district and the creation of a model of the geological structure of the study area. Gammaray spectrometry data analysis is conducted for detailed mapping of intrusive complexes and study of their lithological heterogeneity. Findings. The processing and interpretation of the materials of the conducted field studies in combination with the results of the geologicalgeophysical data analysis made it possible to refine the geological structure and to present a model of oreprospecting complexes in the study area. Areas with an anomalous distribution of potassium, uranium, and thorium have been identified, which make it possible to assess the structural heterogeneity of hidden magmatic massifs and to reveal a connection with gold, copperpolymetallic and raremetal mineralization. Recommendations for further detailed geological exploration, including prospectingexploration drilling, are given. Originality. Research has shown the high efficiency of airborne geophysical methods, such as aeromagnetic and airborne gammaray spectrometry surveys to study the geological structure of ore areas, and in combination with detailed gravimetric exploration to identify lithological heterogeneity of igneous and sedimentary rocks in the study area. New scientificmethodological techniques and the geological structure model based on them gave an opportunity to assess the prospect of the study area for ore mineralization.

В статье рассматривается технология вскрытия и разработки скважин подземного выщелачивания урана. Исследование направлено на повышение эффективности скважинной добычи урана за счет выбора рациональных параметров бурения и освоения призабойной зоны продуктивного горизонта с учетом гидрогеологических характеристик. Рассмотрены параметры бурения с обратной промывкой, свойства руд и промывочных растворов, а также условия подготовки геотехнологических скважин к эксплуатации.

В статье рассматривается построение трехмерной модели Акбулакского редкоземельного месторождения Северного Казахстана. Исследование связано с применением современных геоинформационных технологий, включая MapInfo, ArcGIS, GIS-Micromine, для создания 3D-моделей месторождений минерального сырья. На основе графических и расчетных данных построены каркасная и блочная модели месторождения, а также модели распределения рудных тел по разведочным профилям.