Fast 3D Imaging System Detects & Maps Gas Leaks | TechXplore

New 3D Imaging System Offers Rapid Gas Leak Detection

Researchers at the Anhui Institute of Optics and Fine Mechanics, part of the Hefei Institutes of Physical Science, Chinese Academy of Sciences, have developed a fast, multi-platform compatible system capable of detecting gas leaks in three dimensions. The new technology addresses a critical limitation of existing gas detection methods, which typically provide only two-dimensional projections, hindering accurate assessment of leak volume, distribution, and source.

The development, detailed in recent publications in Environment International and Remote Sensing, focuses on three key areas: gas species identification, precise leak localization, and accurate volume quantification. The team’s approach utilizes both a compact multispectral imaging system for rapid leak scenarios and a larger-scale imager coupled with a deep learning network for broader plume detection.

Addressing the Limitations of 2D Gas Detection

Traditional gas leak detection relies heavily on remote sensing, but these systems often fall short when it comes to providing the detailed 3D information needed for effective response and mitigation. Understanding the volume and distribution of a gas leak is crucial for assessing the potential for fire, explosion, and environmental impact. Without this information, emergency responders and industrial safety teams are operating with incomplete data.

Multispectral Imaging for Rapid Response

For quickly developing leaks, the researchers built a multispectral imaging system integrating infrared detectors, lenses, and motorized components. This system leverages the YOLOv10 model to achieve real-time detection at a rate exceeding 25 frames per second. Crucially, the system doesn’t just identify the presence of a leak; it reconstructs a 3D representation of the gas cloud within 200 milliseconds. Performance metrics, including a Peak Signal-to-Noise Ratio (PSNR) of 25.633 and a Structural Similarity Index (SSIM) of 0.940, demonstrate a high degree of reconstruction accuracy.

The speed and accuracy of this system are enabled by a non-axisymmetric inverse Abel transform, a mathematical technique used to reconstruct 3D data from 2D projections. This allows for a rapid and reliable assessment of the leak’s characteristics, even in dynamic situations.

Large-Scale Plume Mapping with Deep Learning

Recognizing the need for broader coverage, the team also developed the ZK-FTIR-GS1000 imager, paired with a deep learning-based 3D reconstruction network. This network employs an octree representation – a hierarchical data structure – to reconstruct gas clouds efficiently, starting with a coarse resolution and refining it to a finer level of detail with minimal computational demand. Field tests have confirmed the system’s ability to accurately capture the spatial location and distribution of gas leaks, making it well-suited for large-scale environmental monitoring and industrial safety applications.

Implications for Safety and Environmental Monitoring

The development of these tools represents a significant advancement in gas leak detection technology. By providing rapid and accurate 3D reconstructions of gas plumes, the system offers substantial benefits for a range of applications. These include improved environmental monitoring, faster and more effective emergency response, and enhanced industrial safety protocols. The ability to quickly assess the volume and distribution of a leak allows for more targeted mitigation efforts, reducing the risk of escalation and minimizing environmental damage.

The researchers emphasize that this work provides “robust tools for rapid 3D detection and spatial reconstruction of gas leaks, offering strong technical support for environmental monitoring, emergency response, and industrial safety.” The combination of advanced imaging techniques and deep learning algorithms promises to significantly improve the safety and efficiency of gas leak detection in a variety of settings.

More information:
Lei Zhang et al, Spatial location and distribution reconstruction of the leaking gas plume via a single infrared remote sensing system, Environment International (2026). DOI: 10.1016/j.envint.2026.110061
Lei Zhang et al, A 3D Reconstruction of Gas Cloud Leakage Based on Multi-Spectral Imaging Systems, Remote Sensing (2025). DOI: 10.3390/rs17101786