News Categories

How can hyperspectral remote sensing locate hydrocarbon microleakage? — From principles to case studies
Hydrocarbon microleakage is the slow infiltration of light hydrocarbons such as methane and ethane to the near-surface, which is difficult to observe directly with the naked eye. Hyperspectral remote sensing relies on capturing three types of surface traces for detection: the characteristic absorption of hydrocarbon molecules themselves in short-wave infrared light, the mineral alteration spectrum caused by leakage, and the red-edge blue shift caused by vegetation stress due to changes in the soil environment. By extracting hydrocarbon absorption indices and REP data of vegetation red-edge locations and performing spatial overlay analysis, potential oil and gas target areas can be delineated, providing priority field verification clues for oil and gas exploration.

From Spectral Matching to Fundamental Large Models: Deciphering Hyperspectral Target Detection Technology
Hyperspectral target detection is a remote sensing technique that uses known prior spectra of targets to locate their positions within complex backgrounds in hyperspectral images. The process involves acquiring the target's prior spectrum, calculating the response through pixel-by-pixel comparison, generating a detection response map, and obtaining the target result through threshold segmentation. This technique faces challenges such as spectral distortion, mixed pixels, and small sample sizes. The technological approach has evolved from early statistical models and sparse representations to deep learning and large-scale hyperspectral fundamental models.

Methane Spectral Reconstruction Based on FP Interferometer: From Interference Signal to Gas Spectrum
This paper introduces the principle of methane spectral remote sensing based on the Fabry-Perot (FP) interferometer: the FP interferometer acts as a spectral encoder, modulating the methane spectrum into an interference signal; then, using algorithms such as transfer matrix and TSVD regularization, the spectrum is reconstructed to restore the characteristic absorption spectrum of methane. This scheme has a compact structure and can be applied to the monitoring of methane emission sources such as oil and gas and coal mines, providing a new approach for hyperspectral gas remote sensing.

From "Water Injection Reports" to "Integrated Space-Ground Communication": How Hyperspectral Satellites Are Reshaping Ecological and Environmental Monitoring
CCTV exposed multiple cases of falsified environmental monitoring processes in various regions, revealing loopholes in traditional ground-based manual sampling and monitoring that could be manipulated by humans. With the implementation of regulations such as the "Regulations on Ecological and Environmental Monitoring" and the Ecological and Environmental Code, integrated space-ground remote sensing has gained legal support. Hyperspectral satellites, with their spectral fingerprinting capabilities, can identify material components and pollutant concentrations. The Xiguang series of hyperspectral satellites from the Chinese Academy of Sciences provides objective and traceable space-based monitoring methods for ecological and environmental monitoring, compensating for the shortcomings of ground-based monitoring and preventing data falsification.

On-orbit agricultural application verification of Xiguang-2 01 satellite: Comparative study of crops with Sentinel-2
A comparative experiment on precision agriculture was conducted using hyperspectral imagery from the Xiguang-2 01 satellite and multispectral imagery from the Sentinel-2 satellite. Two vegetation indices, anthocyanin ARI1 and nitrogen MTCI, were selected for inversion verification. The results show that the Xiguang-2 01 satellite can capture subtle spectral changes in the field and identify crop stress and nutrient differentiation within the field, demonstrating superior performance compared to traditional multispectral satellites in refined farmland growth diagnosis and variable fertilization scenarios.

From "Sensing the Sky and Calculating the Earth" to "Calculating the Sky and Calculating the Sky," Unlocking the Value of Hyperspectral Data
Following on from the previous article on hyperspectral data cubes, this article analyzes the pain points of the traditional "space-to-ground computing" model, such as large data volume, slow transmission, and high interpretation latency. It introduces the "Sky-to-Sky Computing" on-board intelligent computing technology system, which relies on hardware reconstruction and a hyperspectral AI software engine. Using the Xi'an Optics-Light 2 01 satellite as a carrier, it directly outputs industry conclusions in orbit, empowering environmental protection, agriculture, mining, emergency response, and remote sensing data assetization.

Multi-source remote sensing data fusion: From "individual efforts" to "collaborative perception"
This article elucidates the necessity of multi-source remote sensing data fusion, introduces a three-layer fusion system at the pixel level, feature level, and decision level, reviews the technological evolution from traditional statistical methods to Transformer deep learning and physically constrained hybrid models, and analyzes existing challenges and the development direction of multimodal remote sensing basic models by combining cases of precision agriculture and forest carbon sink monitoring.

Estimation and Analysis of Plot Clustering and Fragmentation at the Urban, County, and Township Levels in Xi'an
Taking Xi'an City as the study area, this study quantitatively analyzes the spatial differentiation characteristics of land parcel clustering and fragmentation at both the county and township levels, using the clustering degree (AI) and patch density (PD) indices. The results show that the land parcel pattern in Xi'an exhibits high clustering at the periphery and high fragmentation at the core, demonstrating a significant scale effect. This analysis can provide support for land spatial planning, farmland protection, and land consolidation.

Using spectral flaws to expose disguises
This article introduces how hyperspectral imaging overcomes the limitations of visible light camouflage by relying on the intrinsic spectral characteristics of materials. It compares traditional spectral detection algorithms with deep learning methods, introduces the HyperCOD public benchmark dataset and the HSC-SAM and DAHNet models, and elaborates on the principles, progress and existing challenges of hyperspectral camouflage target detection technology.

Mapping and Atlas Integration | Decoding the Hyperspectral Data Cube: From "Seeing the Earth's Surface" to "Understanding the Earth"
Compared with traditional optical and multispectral remote sensing, this paper analyzes the three-dimensional structural principle of hyperspectral data cubes that integrate images and spectra, and introduces its application in refined quantitative remote sensing in agriculture, forestry, mining, and water environment. It also points out the industry pain points of data storage, transmission and processing timeliness under the "sky-sensing-ground-computing" model.

