Retrieval Accuracy Research Articles

Machine learning-based estimation of land surface temperature variability over a large region: a temporally consistent approach using single-day satellite imagery.

Accurate retrieval of LST is crucial for understanding and mitigating the effects of urban heat islands, and ultimately addressing the broader challenge of global warming. This study emphasizes the importance of a single day satellite imageries for large-scale LST retrieval. It explores the impact of Spectral indices of the surface parameters, using machine learning algorithms to enhance accuracy. The research proposes a novel approach of capturing satellite data on a single day to reduce uncertainties in LST estimations. A case study over Chandigarh city using Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine, and Random Forest (RF) reveals RF's superior performance in LST estimations during both summer and winter seasons. All the ML models gave an R-square of above 0.8 and RF with slightly higher R-square during both summer (0.93) and winter (0.85). Building on these findings, the study extends its focus to Ranchi, demonstrating RF's robustness with impressive accuracy in capturing LST variations. The research contributes to bridging existing gaps in large-scale LST estimation methodologies, offering valuable insights for its diverse applications in understanding Earth's dynamic systems.

Comparisons of aerosol types and optical characters over Shouxian Area China observed from ground- and space-based systems.

This study evaluates the performance of moderate-resolution Imaging spectroradiometer (MODIS) in aerosol optical depth(AOD) and Ångström exponent(AE) retrievals under high aerosol loading conditions across various aerosol types, utilizing ground-based and space-borne aerosol measurements in Shouxian, China. The intercomparison reveals cloud-aerosol LiDAR with orthogonal polarization's (CALIOP) efficacy in detecting significant aerosol layers and the refinement of sunphotometer-based aerosol type classification through CALIPSO, achieving approximately 80% accuracy. Analysis of 2016-2017 data indicates substantial aerosol presence in Shouxian, with monthly mean AODs ranging from 0.35 to 0.72 at 550 nm, significantly above the global average. The predominant aerosol types were mixed-type (54.8%), desert dust (21.2%), urban/industrial(15.5%), biomass-burning aerosol (6.4%), and continental aerosol (12.1%), with frequent observations of elevated long-range transported aerosol layers. MODIS AOD retrievals generally align with sunphotometer measurements but exhibit higher biases, especially with increasing AOD magnitudes. However, there is a notable difference between MODIS and sunphotometer aerosol AE measurements, with MODIS accurately assessing BBA but showing varied performance across other aerosol types. The combination of AOD and AE of the DD aerosol type is the most accurate. Further analysis showed that MODIS AOD biases and AE biases are negatively correlated, these negative bias correlations show strong aerosol type sensitivities. Monthly analysis of MODIS and sunphotometer comparisons highlights varying performance, particularly during normalized difference vegetation index (NDVI) transitions, suggesting that local vegetation cycles and associated surface spectral reflectance changes significantly impact MODIS aerosol retrieval accuracy under high aerosol loading conditions.

Channel selection method for the CH4 profile retrieval using the Atmospheric Sounder Spectrometer by Infrared Spectral Technology

Methane (CH4) is the second-largest greenhouse gas contributing to global warming, surpassed only by CO2, has a large difference in its vertical concentration distribution, and closely affects the global environment and climate change. The variations in the vertical concentrations of CH4 need to be monitored. Ground-based infrared hyperspectrometers can measure the fine variations of the CH4 concentrations in the vertical distribution within the planetary boundary layer (PBL). However, different detection channels are easily affected by instrumental noise and other environmental factors, leading to differences in the channel spectral characteristics and thereby affecting the accuracy of the CH4 profile retrieval. In this study, an information-weighted channel selection method is proposed for the CH4 profile retrieval using the Atmospheric Sounder Spectrometer by Infrared Spectral Technology (ASSIST) to address the differences in the channel characteristics from the different interference factors. This method leverages the information content of CH4 and its environmental interference factors in each channel to derive the weighting factors, and a comprehensive weighting approach is subsequently applied to ascertain the effective information content of CH4. The method then establishes the threshold for the effective CH4 information content, considering the influence of noise, to select the optimal channels. We employ this method in our study, and 22 channels are selected as the optimal channels for the CH4 profile retrieval. We also evaluate the retrieval capability of the CH4 profile and the anti-interference ability of the selected channels using simulated spectra under clear-sky conditions. When retrieving the CH4 profile using the 1200–1390 cm−1 band (394 channels in total), the CH4 profile is mainly affected by temperature, water vapor, aerosol optical depth (AOD) and N2O. In addition, the mean absolute error (MAE) and the root mean square error (RMSE) for the CH4 profile retrieval using the selected channels are substantially reduced.

Estimating effects of wind and waves on the Doppler centroid frequency shift for the SAR retrieval of ocean currents

Estimating the effects of wind and waves (EWW) on the Doppler centroid frequency shift is a key step in the Synthetic Aperture Radar (SAR) retrieval of ocean currents. However, available models for estimating EWW may exhibit errors as they fail to fully consider the wave and current effects or uncorrected electromagnetic pointing errors. In this study, we corrected the electromagnetic pointing error of Sentinel-1A. In particular, by matching the current data provided by the high-frequency (HF) radar, we accurately obtained the EWW from July 2020 to June 2023. By using the wind and wave parameters provided by the European Centre for Medium-Range Weather Forecasts (ECMWF), the variation characteristics of EWW were analyzed and four geophysical model functions (GMFs) were established using different wave parameters. The application of these GMFs to the retrieval of currents was validated using the HF radar. The optimal current retrieval accuracy in terms of the time series and statistical results was then achieved using the model containing the most comprehensive wave parameters, with a bias of 0.01 m/s and a root mean square error of 0.19 m/s. The proposed model demonstrates the ability to retrieve 100 km scale eddy currents in the Gulf Stream, exhibiting a high agreement with HYbrid Coordinate Ocean Model (HYCOM) reanalysis data, with a bias and root mean square error of approximately −0.03–0.13 m/s and 0.2–0.32 m/s, respectively. The results fully demonstrate that the GMFs established in this study can facilitate the improvement of the SAR retrievals of ocean currents. Moreover, the ability of the proposed model to capture the velocity variations of subscale eddy currents is proved, providing an important tool for oceanographic research.

Research on GNSS-IR soil moisture retrieval based on random forest algorithm

Abstract Soil moisture (SM) retrieval is of great significance in climate, agriculture, ecology, hydrology, and natural disaster monitoring, and it is one of the essential hydrometeorological parameters studied in the world at present. With the continuous development of the GNSS, a technique called GNSS-IR became widely used in ground SM inversion. Therefore, based on the frequency, amplitude and phase of signal-to-noise ratio residuals (δSNR), this study takes P037 and P043 stations set by UNAVCO in the United States as examples and develops the research of SM inversion from Random Forest Regression (RFR) prediction. The experimental results show that the retrieval accuracy of SM under different practical schemes can be in descending order: L1 + L2 dual frequency combination > L2 single frequency > L1 single frequency. It is confirmed that the experimental scheme based on the L1+L2 dual-frequency combination is beneficial to the inversion of SM. In the L1+L2 dual-frequency combination, the prediction set accuracy of the P037 station is as follows: R is 0.796, RMSE is 0.032 cm3cm-3, ME is 0.002 cm3cm-3. The prediction accuracy of the P043 station is as follows: R is 0.858, RMSE is 0.039 cm3cm-3, ME is -0.009 cm3cm-3. Among them, the RMSE of the L1+L2 dual-frequency combination of the two stations has an improvement effect of 13%-37% compared with their single-frequency, which has a noticeable improvement effect. The difference between the SM retrieved by GNSS-IR and the reference value of PBO-H2O is concentrated around 0, further showing the accuracy of SM retrieved by GNSS-IR technology. To sum up, this study considers that SM retrieval based on the RFR model has good reliability and accuracy, which makes GNSS-IR technology an efficient means for SM retrieval. With the continuous improvement of the GNSS system and technology, the application of GNSS-IR technology in SM will become broader.

StarDICE II: Calibration of an Uncooled Infrared Thermal Camera for Atmospheric Gray Extinction Characterization.

The StarDICE experiment strives to establish an instrumental metrology chain with a targeted accuracy of 1 mmag in griz bandpasses to meet the calibration requirements of next-generation cosmological surveys. Atmospheric transmission is a significant source of systematic uncertainty. We propose a solution relying on an uncooled infrared thermal camera to evaluate gray extinction variations. However, achieving accurate measurements with thermal imaging systems necessitates prior calibration due to temperature-induced effects, compromising their spatial and temporal precision. Moreover, these systems cannot provide scene radiance in physical units by default. This study introduces a new calibration process utilizing a tailored forward modeling approach. The method incorporates sensor, housing, flat-field support, and ambient temperatures, along with raw digital response, as input data. Experimental measurements were conducted inside a climatic chamber, with a FLIR Tau2 camera imaging a thermoregulated blackbody source. The results demonstrate the calibration effectiveness, achieving precise radiance measurements with a temporal pixel dispersion of 0.09 W m-2 sr-1 and residual spatial noise of 0.03 W m-2 sr-1. We emphasize that the accuracy of scene radiance retrieval can be systematically affected by the camera's close thermal environment, especially when the ambient temperature exceeds that of the scene.

Benchmarking PathCLIP for Pathology Image Analysis.

Accurate image classification and retrieval are of importance for clinical diagnosis and treatment decision-making. The recent contrastive language-image pre-training (CLIP) model has shown remarkable proficiency in understanding natural images. Drawing inspiration from CLIP, pathology-dedicated CLIP (PathCLIP) has been developed, utilizing over 200,000 image and text pairs in training. While the performance the PathCLIP is impressive, its robustness under a wide range of image corruptions remains unknown. Therefore, we conduct an extensive evaluation to analyze the performance of PathCLIP on various corrupted images from the datasets of osteosarcoma and WSSS4LUAD. In our experiments, we introduce eleven corruption types including brightness, contrast, defocus, resolution, saturation, hue, markup, deformation, incompleteness, rotation, and flipping at various settings. Through experiments, we find that PathCLIP surpasses OpenAI-CLIP and the pathology language-image pre-training (PLIP) model in zero-shot classification. It is relatively robust to image corruptions including contrast, saturation, incompleteness, and orientation factors. Among the eleven corruptions, hue, markup, deformation, defocus, and resolution can cause relatively severe performance fluctuation of the PathCLIP. This indicates that ensuring the quality of images is crucial before conducting a clinical test. Additionally, we assess the robustness of PathCLIP in the task of image-to-image retrieval, revealing that PathCLIP performs less effectively than PLIP on osteosarcoma but performs better on WSSS4LUAD under diverse corruptions. Overall, PathCLIP presents impressive zero-shot classification and retrieval performance for pathology images, but appropriate care needs to be taken when using it.

Fast retrieval of XCO2 over east Asia based on Orbiting Carbon Observatory-2 (OCO-2) spectral measurements

Abstract. The increase in greenhouse gas concentrations, particularly CO2, has significant implications for global climate patterns and various aspects of human life. Spaceborne remote sensing satellites play a crucial role in high-resolution monitoring of atmospheric CO2. However, the next generation of greenhouse gas monitoring satellites is expected to face challenges, particularly in terms of computational efficiency in atmospheric CO2 retrieval and analysis. To address these challenges, this study focuses on improving the speed of retrieving the column-averaged dry-air mole fraction of carbon dioxide (XCO2) using spectral data from the Orbiting Carbon Observatory-2 (OCO-2) satellite while still maintaining retrieval accuracy. A novel approach based on neural network (NN) models is proposed to tackle the nonlinear inversion problems associated with XCO2 retrievals. The study employs a data-driven supervised learning method and explores two distinct training strategies. Firstly, training is conducted using experimental data obtained from the inversion of the operational optimization model, which is released as the OCO-2 satellite products. Secondly, training is performed using a simulated dataset generated by an accurate forward calculation model. The inversion performance and prediction performance of the machine learning model for XCO2 are compared, analyzed, and discussed for the observed region over east Asia. The results demonstrate that the model trained on simulated data accurately predicts XCO2 in the target area. Furthermore, when compared to OCO-2 satellite product data, the developed XCO2 retrieval model not only achieves rapid predictions (<1 ms) with good accuracy (1.8 ppm or approximately 0.45 %) but also effectively captures sudden increases in XCO2 plumes near industrial emission sources. The accuracy of the machine learning model retrieval results is validated against reliable data from Total Carbon Column Observing Network (TCCON) sites, demonstrating its ability to effectively capture CO2 seasonal variations and annual growth trends.

Deep Attention Fusion Hashing (DAFH) Model for Medical Image Retrieval.

To address this, we propose a novel deep learning-based hashing model, the Deep Attention Fusion Hashing (DAFH) model, which integrates advanced attention mechanisms with medical imaging data. The DAFH model enhances retrieval performance by integrating multi-modality medical imaging data and employing attention mechanisms to optimize the feature extraction process. Utilizing multimodal medical image data from the Cancer Imaging Archive (TCIA), this study constructed and trained a deep hashing network that achieves high-precision classification of various cancer types. At hash code lengths of 16, 32, and 48 bits, the model respectively attained Mean Average Precision (MAP@10) values of 0.711, 0.754, and 0.762, highlighting the potential and advantage of the DAFH model in medical image retrieval. The DAFH model demonstrates significant improvements in the efficiency and accuracy of medical image retrieval, proving to be a valuable tool in clinical settings.

Towards long-term, high-accuracy, and continuous satellite total and fine-mode aerosol records: Enhanced Land General Aerosol (e-LaGA) retrieval algorithm for VIIRS

Long-term, accurate, stable, and continuous aerosol records from space are a major requirement for climate and atmospheric environment research. Due to the limited period span of the single satellite mission, solving the problem requires the combination of multiple satellite missions. In this study, a novel aerosol retrieval algorithm, named enhanced Land General Aerosol Retrieval (e-LaGA), for Visible/Infrared Imager Radiometer Suite (VIIRS) was developed to continue Moderate-resolution Imaging Spectro-radiometer (MODIS) aerosol retrieval. e-LaGA utilizes a Surface Reflectance (SR) relationship model map to more accurately estimate the SR parameter, optimizes the priori aerosol-type map, and improves the multi-band retrieval strategy using the residual-interpolation approach. Additionally, e-LaGA expands the retrieval ability of the traditional Dark Target (DT) algorithm over bright surfaces, and can more accurately retrieve Aerosol Optical Depth (AOD), Fine-mode AOD (AODF), and Fraction (FMF). The validation using global 533 AERONET sites shows that the volume of matchups of AOD (AODF) is approximately 80000, the correlation coefficient (R) is 0.902 (0.879) and the fraction of meeting the expected error envelope (±0.05 ± 0.15τ) is 0.806 (0.804). The inter-comparison indicates that the accuracy of e-LaGA AOD retrievals is comparable to seven commonly used AOD products. The validation performance and spatial distribution pattern of e-LaGA AODF and FMF are in good agreement with the POLDER (Polarization and Directionality of the Earth's Reflectances) products. e-LaGA is also applied to MODIS sensors. The VIIRS e-LaGA AOD, AODF, and FMF retrievals have high consistency with MODIS e-LaGA. Their average bias of AOD is 0.006, which is smaller than 0.024 for the Deep Blue and 0.011 for the DT. These preliminary results demonstrate the robustness of the e-LaGA algorithm and its potential for establishing a long-term climate record of total and fine-mode aerosol by combining multiple satellite missions, which is expected to reduce the uncertainty of climate change research.

Dominant activities of fear engram cells in the dorsal dentate gyrus underlie fear generalization in mice.

Over-generalized fear is a maladaptive response to harmless stimuli or situations characteristic of posttraumatic stress disorder (PTSD) and other anxiety disorders. The dorsal dentate gyrus (dDG) contains engram cells that play a crucial role in accurate memory retrieval. However, the coordination mechanism of neuronal subpopulations within the dDG network during fear generalization is not well understood. Here, with the Tet-off system combined with immunostaining and two-photon calcium imaging, we report that dDG fear engram cells labeled in the conditioned context constitutes a significantly higher proportion of dDG neurons activated in a similar context where mice show generalized fear. The activation of these dDG fear engram cells encoding the conditioned context is both sufficient and necessary for inducing fear generalization in the similar context. Activities of mossy cells in the ventral dentate gyrus (vMCs) are significantly suppressed in mice showing fear generalization in a similar context, and activating the vMCs-dDG pathway suppresses generalized but not conditioned fear. Finally, modifying fear memory engrams in the dDG with "safety" signals effectively rescues fear generalization. These findings reveal that the competitive advantage of dDG engram cells underlies fear generalization, which can be rescued by activating the vMCs-dDG pathway or modifying fear memory engrams, and provide novel insights into the dDG network as the neuronal basis of fear generalization.

SWIRL: The First Australian Operational Radar-Based 3D Wind Analysis System

Abstract In this paper, the first Australian operational radar-based three-dimensional (3D) wind analysis system named Synthetic Wind Information from Radar and Lidar (SWIRL) is described and evaluated. SWIRL employs a variational minimization formulation to combine results from four individual wind retrieval techniques of varied complexity to derive 3D winds in single-Doppler and multi-Doppler radar regions: a variational version of the traditional velocity azimuth display (VVAD) and double VAD (DVAD) techniques, a single-Doppler wind retrieval technique using optical flow horizontal wind proxies, and a multi-Doppler 3D wind retrieval technique. The SWIRL 3D wind components are evaluated against wind profiler observations and radar simulations using a very high-resolution (50 m) numerical simulation of a supercell thunderstorm. We find that SWIRL can retrieve very accurate horizontal winds, especially below 2-km height in the multi-Doppler regions, with mean absolute errors on wind speed and direction < 2 m s−1 and 10° on average and <2.5 m s−1 and 15°–20° 90% of the time. These errors do not increase noticeably with wind speed, highlighting the suitability of these retrieved winds to be used for damaging and destructive wind detection and nowcasting. The single-Doppler retrieval using optical flow is also found to provide reasonably accurate winds at these heights. The accurate retrieval of convective-scale updrafts and downdrafts, even using multi-Doppler information, is still a major challenge, with mean absolute errors of vertical velocity of about 50% on average. This can be attributed to the limitations of the current radar technology used operationally, imposing slow antenna speeds. Significance Statement Damaging and destructive winds have the potential to inflict significant damage to properties and assets and, tragically, result in loss of life. Efficient direction of emergency services to affected areas is essential for a prompt return to normal conditions. Wind farm operators require precise information on anticipated wind shifts to reduce the risk of energy grid failures. Strong winds also contribute to compound weather events, such as water ingress through hail-damaged roofs or structural damage to buildings caused by hailstones. The purpose of this work was to equip Australia with the first operational wind monitoring system, based on operational radar observations, to serve all these critical applications (and more).

Striving towards robust phase diversity on-sky

Context. The recent IRLOS upgrade for VLT/MUSE narrow field mode (NFM) introduced a full-pupil mode to enhance sensitivity and sky coverage. This involved replacing the 2 × 2 Shack-Hartmann sensor with a single lens for full-aperture photon collection, which also enabled the engagement of the linearized focal-plane technique (LIFT) wavefront sensor instead. However, initial on-sky LIFT experiments have highlighted a complex point spread function (PSF) structure due to strong and polychromatic non-common path aberrations (NCPAs), complicating the accurate retrieval of tip-tilt and focus using LIFT. Aims. This study aims to conduct the first on-sky validation of LIFT on VLT/UT4, outline challenges encountered during the tests, and propose solutions for increasing the robustness of LIFT in on-sky operations. Methods. We developed a two-stage approach to focal-plane wavefront sensing, where tip-tilt and focus retrieval done with LIFT is preceded by the NCPA calibration step. The resulting NCPA estimate is subsequently used by LIFT. To perform the calibration, we proposed a method capable of retrieving the information about NCPAs directly from on-sky focal-plane PSFs. Results. We verified the efficacy of this approach in simulated and on-sky tests. Our results demonstrate that adopting the two-stage approach has led to a significant improvement in the accuracy of the defocus estimation performed by LIFT, even under challenging low-flux conditions. Conclusions. The efficacy of LIFT as a slow and truth focus sensor in practical scenarios has been demonstrated. However, integrating NCPA calibration with LIFT is essential to verifying its practical application in the real system. Additionally, the proposed calibration step can serve as an independent and minimally invasive approach to evaluate NCPA on-sky.

Land Surface Longwave Radiation Retrieval from ASTER Clear-Sky Observations

Surface longwave radiation (SLR) plays a pivotal role in the Earth’s energy balance, influencing a range of environmental processes and climate dynamics. As the demand for high spatial resolution remote sensing products grows, there is an increasing need for accurate SLR retrieval with enhanced spatial detail. This study focuses on the development and validation of models to estimate SLR using measurements from the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) sensor. Given the limitations posed by fewer spectral bands and data products in ASTER compared to moderate-resolution sensors, the proposed approach combines an atmospheric radiative transfer model MODerate resolution atmospheric TRANsmission (MODTRAN) with the Light Gradient Boosting Machine algorithm to estimate SLR. The MODTRAN simulations were performed to construct a representative training dataset based on comprehensive global atmospheric profiles and surface emissivity spectra data. Global sensitivity analyses reveal that key inputs influencing the accuracy of SLR retrievals should reflect surface thermal radiative signals and near-surface atmospheric conditions. Validated against ground-based measurements, surface upward longwave radiation (SULR) and surface downward longwave radiation (SDLR) using ASTER thermal infrared bands and surface elevation estimations resulted in root mean square errors of 17.76 W/m2 and 25.36 W/m2, with biases of 3.42 W/m2 and 3.92 W/m2, respectively. Retrievals show systematic biases related to extreme temperature and moisture conditions, e.g., causing overestimation of SULR in hot humid conditions and underestimation of SDLR in arid conditions. While challenges persist, particularly in addressing atmospheric variables and cloud masking, this work lays a foundation for accurate SLR retrieval from high spatial resolution sensors like ASTER. The potential applications extend to upcoming satellite missions, such as the Landsat Next, and contribute to advancing high-resolution remote sensing capabilities for an improved understanding of Earth’s energy dynamics.

Outcome Prediction Using Multi-Modal Information: Integrating Large Language Model-Extracted Clinical Information and Image Analysis.

Survival prediction post-cystectomy is essential for the follow-up care of bladder cancer patients. This study aimed to evaluate artificial intelligence (AI)-large language models (LLMs) for extracting clinical information and improving image analysis, with an initial application involving predicting five-year survival rates of patients after radical cystectomy for bladder cancer. Data were retrospectively collected from medical records and CT urograms (CTUs) of bladder cancer patients between 2001 and 2020. Of 781 patients, 163 underwent chemotherapy, had pre- and post-chemotherapy CTUs, underwent radical cystectomy, and had an available post-surgery five-year survival follow-up. Five AI-LLMs (Dolly-v2, Vicuna-13b, Llama-2.0-13b, GPT-3.5, and GPT-4.0) were used to extract clinical descriptors from each patient's medical records. As a reference standard, clinical descriptors were also extracted manually. Radiomics and deep learning descriptors were extracted from CTU images. The developed multi-modal predictive model, CRD, was based on the clinical (C), radiomics (R), and deep learning (D) descriptors. The LLM retrieval accuracy was assessed. The performances of the survival predictive models were evaluated using AUC and Kaplan-Meier analysis. For the 163 patients (mean age 64 ± 9 years; M:F 131:32), the LLMs achieved extraction accuracies of 74%~87% (Dolly), 76%~83% (Vicuna), 82%~93% (Llama), 85%~91% (GPT-3.5), and 94%~97% (GPT-4.0). For a test dataset of 64 patients, the CRD model achieved AUCs of 0.89 ± 0.04 (manually extracted information), 0.87 ± 0.05 (Dolly), 0.83 ± 0.06~0.84 ± 0.05 (Vicuna), 0.81 ± 0.06~0.86 ± 0.05 (Llama), 0.85 ± 0.05~0.88 ± 0.05 (GPT-3.5), and 0.87 ± 0.05~0.88 ± 0.05 (GPT-4.0). This study demonstrates the use of LLM model-extracted clinical information, in conjunction with imaging analysis, to improve the prediction of clinical outcomes, with bladder cancer as an initial example.

Intelligent Text Mining for Ontological Knowledge Graph Refinement and Patent Portfolio Analysis—Case Study of Net-Zero Data Center Innovation Management

Ontological knowledge graph (OKG) is a well-formed visual representation that depicts knowledge organization in formal elements (e.g., entities and attributes) and their interrelationships. OKG is crucial for innovation management analysis as it provides a clear boundary to understand complex knowledge domain in detail. In the patent analysis field, it facilitates the definition of a well-defined patent portfolio, aiming for accurate and complete patent retrievals and subsequent analyses. In recent decade, the rapid growth of the information and communication technology (ICT) sector has rendered data centers (DCs) indispensable for data processing, storage, and cloud computing, while ensuring security and privacy during DC operations. However, their energy-intensive operations pose challenges to global efforts toward achieving net-zero emissions goals. In response, this research develops a formal OKG refinement process and uses DC net-zero technology OKG as case study for in-depth OKG refinement and application in patent portfolio analysis. The net-zero DC domain covers five sub-technologies. Utilizing the proposed OKG refinement and patent portfolio analysis framework, the 1801 most recent decade’s patents related to relevant “DC net-zero technologies” are retrieved and analyzed. Particularly in this case, DC colocation and server-as-a-service perspectives are the newly discovered sub-domains for OKG refinement. Furthermore, the research also adopts the technology function matrix and technology maturity to assess current and future technology development trends, providing crucial insights supporting strategic innovation management.

Development of the digital retrieval system integrating intelligent information and improved genetic algorithm: A study based on art museums.

This study aims to develop a digital retrieval system for art museums to solve the problems of inaccurate information and low retrieval efficiency in the digital management of cultural heritage. By introducing an improved Genetic Algorithm (GA), digital management and access efficiency are enhanced, to bring substantial optimization and innovation to the digital management of cultural heritage. Based on the collection of art museums, this study first integrates the collection's images, texts, and metadata with multi-source intelligent information to achieve a more accurate and comprehensive description of digital content. Second, a GA is introduced, and a GA 2 Convolutional Neural Network (GA2CNN) optimization model combining domain knowledge is proposed. Moreover, the convergence speed of traditional GA is improved to adapt to the characteristics of cultural heritage data. Lastly, the Convolutional Neural Network (CNN), GA, and GA2CNN are compared to verify the proposed system's superiority. The results show that in all models, the sample output results' actual value is 2.62, which represents the real data observation results. For sample number 5, compared with the actual value of 2.62, the predicted values of the GA2CNN and GA models are 2.6177 and 2.6313, and their errors are 0.0023 and 0.0113. The CNN model's predicted value is 2.6237, with an error of 0.0037. It can be found that the network fitting accuracy after optimization of the GA2CNN model is high, and the predicted value is very close to the actual value. The digital retrieval system integrated with the GA2CNN model has a good performance in enhancing retrieval efficiency and accuracy. This study provides technical support for the digital organization and display of cultural heritage and offers valuable references for innovative exploration of museum information management in the digital era.

Inversion of Farmland Soil Moisture Based on Multi-Band Synthetic Aperture Radar Data and Optical Data

Surface soil moisture (SSM) plays an important role in agricultural and environmental systems. With the continuous improvement in the availability of remote sensing data, satellite technology has experienced widespread development in the monitoring of large-scale SSM. Synthetic Aperture Radar (SAR) and optical remote sensing data have been extensively utilized due to their complementary advantages in this field. However, the limited information from single-band SARs or single optical remote sensing data has restricted the accuracy of SSM retrieval, posing challenges for precise SSM monitoring. In contrast, multi-source and multi-band remote sensing data contain richer and more comprehensive surface information. Therefore, a method of combining multi-band SAR data and employing machine learning models for SSM inversion was proposed. C-band Sentinel-1 SAR data, X-band TerraSAR data, and Sentinel-2 optical data were used in this study. Six commonly used feature parameters were extracted from these data. Three machine learning methods suitable for small-sample training, including Genetic Algorithms Back Propagation (GA-BP), support vector regression (SVR), and Random Forest (RF), were employed to construct the SSM inversion models. The differences in SSM retrieval accuracy were compared when two different bands of SAR data were combined with optical data separately and when three types of data were used together. The results show that the best inversion performance was achieved when all three types of remote sensing data were used simultaneously. Additionally, compared to the C-band SAR data, the X-band SAR data exhibited superior performance. Ultimately, the RF model achieved the best accuracy, with a determinable coefficient of 0.9186, a root mean square error of 0.0153 cm3/cm3, and a mean absolute error of 0.0122 cm3/cm3. The results indicate that utilizing multi-band remote sensing data for SSM inversion offers significant advantages, providing a new perspective for the precise monitoring of SSM.

Multi-FusNet: fusion mapping of features for fine-grained image retrieval networks.

As the diversity and volume of images continue to grow, the demand for efficient fine-grained image retrieval has surged across numerous fields. However, the current deep learning-based approaches to fine-grained image retrieval often concentrate solely on the top-layer features, neglecting the relevant information carried in the middle layer, even though these information contains more fine-grained identification content. Moreover, these methods typically employ a uniform weighting strategy during hash code mapping, risking the loss of critical region mapping-an irreversible detriment to fine-grained retrieval tasks. To address the above problems, we propose a novel method for fine-grained image retrieval that leverage feature fusion and hash mapping techniques. Our approach harnesses a multi-level feature cascade, emphasizing not just top-layer but also intermediate-layer image features, and integrates a feature fusion module at each level to enhance the extraction of discriminative information. In addition, we introduce an agent self-attention architecture, marking its first application in this context, which steers the model to prioritize on long-range features, further avoiding the loss of critical regions of the mapping. Finally, our proposed model significantly outperforms existing state-of-the-art, improving the retrieval accuracy by an average of 40% for the 12-bit dataset, 22% for the 24-bit dataset, 16% for the 32-bit dataset, and 11% for the 48-bit dataset across five publicly available fine-grained datasets. We also validate the generalization ability and performance stability of our proposed method by another five datasets and statistical significance tests. Our code can be downloaded from https://github.com/BJFU-CS2012/MuiltNet.git.

Differentiable modeling for soil moisture retrieval by unifying deep neural networks and water cloud model

Machine learning has been widely used in high-spatial-resolution surface soil moisture (SSM) retrieval studies, but in recent years, this purely data-driven retrieval method has been controversial due to its lack of physical interpretability and generalization ability. Physical retrieval models based on the theory of radiative transfer equations respect physical laws, but their retrieval accuracy is usually affected by many insufficient accurate inputs, the complex model structure, and parameter adjustment method. In order to explore the retrieval method of unifying these two types of models, in this study, a differentiable model (DM) was constructed to realize the soil moisture retrieval at 10 m resolution based on Sentinel data. The differentiable soil moisture retrieval model takes the water cloud model (WCM) as the skeleton, and united the WCM and neural networks by implementing differentiable programming of the WCM in a machine learning platform. The differentiability makes the retrieval model trained by the gradient descent method the same as the neural network, which allows the retrieval model to be physical while obtaining more accurate retrieval results. Luan River Basin, Shandian River Basin, Maqu, and Lake Tahoe study areas with various land cover types and climate types were selected for model evaluation, and the performances of DM were close to that of the random forest model with Pearson correlation coefficient (R) of 0.747, 0.853, 0.838 and 0.792 in four study areas, respectively. While in the assessment of extrapolation capability of retrieval models, the DM showed its strong generalization ability and retrieval performance that exceeded that of the other retrieval models, with R of 0.786, unbiased root mean square error (ubRMSE) of 5.523 vol% and bias of 0.054 vol%. The DM synthesizes the advantages of both physical and machine learning models while providing high-resolution SSM estimates with acceptable accuracy for the study areas. This study creates favorable conditions for the realization of large-scale soil moisture retrieval with high resolution and high accuracy, and provides new ideas for the combination of machine learning and physical knowledge in other retrieval studies.