Unsupervised clustering of deep image features is often proposed for unlabeled image archives, yet it is usually judged by internal criteria rather than by agreement with expert annotations. In a single domain of texture images, namely full-size rock-core scans, we evaluate unsupervised partitions against depth-registered expert descriptions of 20 wells (nine classes), holding out entire wells. Color statistics, texture descriptors, VGG16, two DINOv2 models and SimCLR encoders pre-trained on the images are combined with k-means, Gaussian mixtures and SCAN, several seeds, and pre-training runs. Across all combinations, the mean class accuracy plateaus at 0.35–0.40 (chance level 0.12), and six color statistics reach this plateau with a Gaussian mixture model (0.37). Apparent gains of in-domain self-supervised features depend on the clustering algorithm and the pre-training run. The plateau is robust to the evaluation protocol, a coarser class scheme and more diverse fitting wells, and persists on color-normalized and grayscale images. Linear probes on the same frozen features reach 0.47–0.54, which locates the bottleneck in the unsupervised objectives. Fitting the clustering transductively on all unlabeled images lifts in-domain features to 0.42–0.48, but this gain does not transfer to unseen wells. The conclusions are specific to this domain.

Background: Lithology is a vital field of study in both geology and the oil and gas sector that focuses on the properties of geological rocks. The primary objectives of lithology to classify rocks, determine their origin, and investigate the conditions of their formation and changes over time. Lithological core examination employ various methods, encompassing both conventional techniques (e.g., visual inspection of the rock samples or microscopic analysis of slides) and modern technologies. Conventional methods of examination require high qualifications and experience, and can be labour-intensive, especially in visual analysis (description of core material). The application of machine learning methods and automated technologies can enhance the efficiency and accuracy of analysis, save time, and provide quick access to information. Aim: To develop lithotypes clustering model on core images using machine learning methods. Materials and methods: The paper discusses an algorithm for clustering lithotypes using K-Means method combined with VGG16, VGG19 and ResNet50 convolutional neural networks to identify key features (similarities and distinctions as determined from photos). Results: The algorithm for clustering lithotypes using K-Means method and convolutional neural networks is developed. The advantages and limitations of the algorithm when working with core images are determined. Results from experiments conducted using an actual dataset are presented. Conclusion: The findings of the study offer important practical insights that can be applied to deep learning methods for core analysis as well as geological research. The application of this approach in geology can be broadened and the analysis of alternative machine learning models and techniques can be strengthened with more investigation.

Background: Ensuring employees occupational safety at the oil and gas industry facilities remains a relevant objective, as far as traditional monitoring methods for the use of Personal Protective Equipment (PPE) are based on manual visual inspections and are susceptible to human factor. Most existing computer vision systems are limited to detecting only a small number of 2–6 classes of PPE categories and to verifying the anatomical consistency between detected protective equipment and the corresponding body parts of employees. Aim: Developing and validating a digital monitoring method for compliance with PPE use requirements based on neural network architectures, integrating algorithms for object detection, human pose estimation, and anatomical matching of PPE elements in real time mode. Materials and methods: A unique dataset of 16352 images (after augmentation) containing 13 object classes, including 6 types of PPE and 6 negative classes was created for training the model. The YOLOv8 model was used for object detection, and HRNet for human pose estimation. A two-tier video stream processing architecture was implemented, combining object tracking (BoT-SORT), spatial and anatomical matching and TensorRT quantization to improve system performance. Results: During the training phase, the YOLOv8 model achieved Precision = 0.98, Recall = 0.97, and F1 = 0.94. When testing the developed system on 16 video files obtained from industrial sites, the system achieved a precision of 96.58%, a recall of 68.48%, and an F1 score of 0.8014. Detecting small objects (gloves) in cropped images improves detection efficiency by 2-3 times compared to full-frame processing. Conclusion: The developed approach, combining the YOLOv8 and HRNet models and an anatomical matching algorithm, provides effective digital monitoring of compliance with PPE requirements in real-world production conditions. The obtained results confirm the potential of its application in the development of intelligent industrial safety monitoring systems at industrial facilities.

Monitoring of oil-contaminated land is a pressing challenge for Kazakhstan. During the Soviet era, oil production at a number of fields relied on the so-called pit method, whereby drilling waste, process fluids, and associated wastewater were collected in purpose-dug earthen pits adjacent to extraction sites, leaving a substantial legacy of oil-contaminated areas requiring remediation and systematic monitoring. Remote sensing combined with deep learning-based image processing enables more accurate segmentation of such territories and facilitates their ongoing monitoring. This paper presents ROSID-HR, a dataset constructed from PlanetScope multispectral imagery designed for the segmentation of small-area oil contamination. In developing ROSID-HR, oil-contaminated sites from our previously published ROSID dataset were used as reference annotations, ensuring consistency between the two datasets. Through systematic experimentation, the optimal PlanetScope channel configuration for separating oil-contaminated areas from shadows and anthropogenic objects was identified as Blue-Red-NIR (bands 2-6-8). To validate the dataset, semantic segmentation was performed using the convolutional DeepLabv3+ model and the transformer-based Mask2Former architecture. Mask2Former achieved an IoU of 69.73 % for the oil class on the test set, exceeding the DeepLabv3+ result of 21.63 % by a factor of more than three. The results confirm the effectiveness and advantages of using PlanetScope data and the Mask2Former architecture together as part of the considered approaches to detecting small areas contaminated with oil.

Air pollution in Almaty, Kazakhstan, poses a critical public health challenge intensified by the city’s basin topography and seasonal thermal inversions that trap anthropogenic emissions. The sparse stationary network (~5 stations for ~2 million inhabitants) lacks the spatial and temporal resolution needed to capture intra-urban variability. We present the design, deployment, and field evaluation of a low-cost distributed Internet of Things (IoT) network of six custom nodes—Winsen ZPHS01B multi-parameter modules with Raspberry Pi Zero 2 W edge units, at an estimated principal-component cost of ~US$100 per node—operated during a winter-spring campaign (February–April 2025) and yielding over 70,000 measurements of PM_2.5, PM₁₀, CO₂, temperature, and relative humidity. The system’s novelty lies in three integrated engineering features: an Active Airflow Stabilization enclosure that decouples sampling from external wind, context-aware adaptive edge filtering that reduces transmitted data volume by ~40%, and a secure Edge-DMZ-Core telemetry pipeline. Node readings were cross-validated against a Qingping Air Monitor Pro with documented traceability to FEM-grade reference analyzers (R² = 0.89–0.95), and city-scale consistency was confirmed against the national monitoring dashboard; the network is therefore characterized as providing internally consistent low-cost observations rather than reference-equivalent concentrations. Daily mean PM_2.5 exceeded the WHO 24 h guideline (15 µg/m³) on 84% of monitored days, with February concentrations (54.4 µg/m³) significantly above March (21.9 µg/m³; p < 0.001). A high PM_2.5/PM₁₀ ratio (~0.96), measured at the physically consistent nodes, together with higher weekend concentrations, points to coal-based residential heating as the most likely dominant source. A coupled WRF-SILAM framework is configured for future model-observation integration. The system offers a reproducible, scalable, and cost-effective template for ambient particulate monitoring in resource-constrained cities with complex terrain.