Comparison between ozonesonde measurements and satellite retrievals over Beijing,China

从2013年开始,作者团队使用自主研发电化学原理臭氧探空仪在华北平原北京地区进行每周一次观测.本研究首次使用2013-2019年期间北京地区臭氧探空数据评估Aqua卫星搭载大气红外探测仪(AIRS)和Aura卫星搭载微波临边探测器(MLS)反演垂直臭氧廓线,并对比臭氧探空,AIRS和Aura卫星搭载臭氧监测仪(OMI)臭氧柱总量结果.尽管臭氧探空与卫星反演垂直臭氧廓线在局部高度处差异较大,但整体来说两者较为接近(相对偏差大多<10%).臭氧探空,AIRS和OMI三种仪器测量臭氧柱总量的年变化特征较为一致,其年均臭氧柱总量分别为351.8±18.4 DU,348.8±19.5 DU和336.9±14.2 DU.后续对国内多站点观测数据分析将有助于进一步理解臭氧探空与卫星反演臭氧资料在不同区域的一致性.

Importance-aware 3D volume visualization for medical content-based image retrieval-a preliminary study

Background A medical content-based image retrieval(CBIR)system is designed to retrieve images from large imaging repositories that are visually similar to a user′s query image.CBIR is widely used in evidence-based diagnosis,teaching,and research.Although the retrieval accuracy has largely improved,there has been limited development toward visualizing important image features that indicate the similarity of retrieved images.Despite the prevalence of 3D volumetric data in medical imaging such as computed tomography(CT),current CBIR systems still rely on 2D cross-sectional views for the visualization of retrieved images.Such 2D visualization requires users to browse through the image stacks to confirm the similarity of the retrieved images and often involves mental reconstruction of 3D information,including the size,shape,and spatial relations of multiple structures.This process is time-consuming and reliant on users'experience.Methods In this study,we proposed an importance-aware 3D volume visualization method.The rendering parameters were automatically optimized to maximize the visibility of important structures that were detected and prioritized in the retrieval process.We then integrated the proposed visualization into a CBIR system,thereby complementing the 2D cross-sectional views for relevance feedback and further analyses.Results Our preliminary results demonstrate that 3D visualization can provide additional information using multimodal positron emission tomography and computed tomography(PETCT)images of a non-small cell lung cancer dataset.

A Visual Indoor Localization Method Based on Efficient Image Retrieval

The task of indoor visual localization, utilizing camera visual information for user pose calculation, was a core component of Augmented Reality (AR) and Simultaneous Localization and Mapping (SLAM). Existing indoor localization technologies generally used scene-specific 3D representations or were trained on specific datasets, making it challenging to balance accuracy and cost when applied to new scenes. Addressing this issue, this paper proposed a universal indoor visual localization method based on efficient image retrieval. Initially, a Multi-Layer Perceptron (MLP) was employed to aggregate features from intermediate layers of a convolutional neural network, obtaining a global representation of the image. This approach ensured accurate and rapid retrieval of reference images. Subsequently, a new mechanism using Random Sample Consensus (RANSAC) was designed to resolve relative pose ambiguity caused by the essential matrix decomposition based on the five-point method. Finally, the absolute pose of the queried user image was computed, thereby achieving indoor user pose estimation. The proposed indoor localization method was characterized by its simplicity, flexibility, and excellent cross-scene generalization. Experimental results demonstrated a positioning error of 0.09 m and 2.14° on the 7Scenes dataset, and 0.15 m and 6.37° on the 12Scenes dataset. These results convincingly illustrated the outstanding performance of the proposed indoor localization method.

Validation of MODIS aerosol retrievals and evaluation of potential cloud contamination in East Asia

MODIS aerosol retrievals onboard Terra/Aqua and ground truth data obtained from AERONET(Aerosol Robtic Network) solar direct radiance measurements are collocated to evaluate the quality of the former in East Asia. AERONET stations in East Asia are separated into two groups according to their locations and the preliminary validation results for each station. The validation results showed that the accuracy of MODIS aerosol retrievals in East Asia is a little worse than that obtained in other regions such as Eastern U.S., Western Europe, Brazil and so on. The primary reason is due to the improper aerosol model used in MODIS aerosol retrieval algorithm, so it is of significance to characterize aerosol properties properly according to long term ground-based remote sensing or other relevant in situ observations in order to improve MODIS retrievals in East Asia. Cloud contamination is proved to be one of large errors, which is demonstrated by the significant relation between MODIS aerosol retrievals versus cloud fraction, as well as notable improvement of linear relation between satellite and ground aerosol data after potential cloud contamination screened. Hence, it is suggested that more stringent clear sky condition be set in use of MODIS aerosol data. It should be pointed out that the improvement might be offset by other error sources in some cases because of complex relation between different errors. Large seasonal variation of surface reflection and uncertainties associated with it result in large intercepts and random error in MODIS aerosol retrievals in northern inland of East Asia. It remains to be a big problem to retrieve aerosols accurately in inland characterized by relatively larger surface reflection than the requirement in MODIS aerosol retrieval algorithm.

A high wind geophysical model fuction for Quik SCAT wind retrievals and application to Typhoon IOKE

The geophysical model function （GMF） describes the relationship between a backscattering and a sea surface wind, and enables a wind vector retrieval from backscattering measurements. It is clear that the GMF plays an important role in an ocean wind vector retrieval. The performance of the existing Ku-band model function QSCAT-1 is considered to be effective at low and moderate wind speed ranges. However, in the conditions of higher wind speeds, the existing algorithms diverge alarmingly, owing to the lack of in situ data required for developing the GMF for the high wind conditions, the QSCAT-1 appears to overestimate the a0, which results in underestimating the wind speeds. Several match-up QuikSCAT and special sensor microwave/imager （SSM/I） wind speed measurements of the typhoons occurring in the west Pacific Ocean are analyzed. The results show that the SSM/I wind exhibits better agreement with the ＂best track＂ analysis wind speed than the QuikSCAT wind retrieved using QSCAT-1. On the basis of this evaluation, a correction of the QSCAT-1 model function for wind speed above 16 m/s is proposed, which uses the collocated SSM/I and QuikSCAT measurements as a training set, and a neural network approach as a multiple nonlinear regression technologytechnology.In order to validate the revised GMF for high winds, the modified GMF was applied to the QuikSCAT observations of Hurricane IOKE. The wind estimated by the QuikSCAT for Typhoon IOKE in 2006 was improved with the maximum wind speed reaching 55 m/s. An error analysis was performed using the wind fields from the Holland model as the surface truth. The results show an improved agreement with the Holland model wind when compared with the wind estimated using the QSCAT-1. However, large bias still existed, indicating that the effects of rain must be considered for further improvement.

Effect of HITRAN Database Improvement on Retrievals of Atmospheric Carbon Dioxide from Reflected Sunlight Spectra in the 1.61-μm Spectral Window

A large number of experimental and theoretical investigations of carbon dioxide （CO 2 ） spectra have been conducted since the most recent update of the High-Resolution Transmission Molecular Absorption （HITRAN） database. To maintain optimal parameters, the HITRAN 2004 CO 2 line list has been completely replaced by HITRAN 2008 data in the near-infrared region from 4300 cm-1 to 7000 cm-1 . To examine the effect of this change on the retrieval of CO 2 vertical column data from reflected sunlight spectra in the 1.61-μm spectral window, synthetic measurements for a given atmospheric state and instrument setup were generated and compared using radiative transfer model with the line-transition parameters from the HITRAN 2004 and 2008 databases. Simulated retrievals were then performed based on the optimal estimation retrieval theory. The results show that large systematic errors in atmospheric CO 2 column retrievals were induced by the differences in the HITRAN laboratory line parameters in the 1.61-μm region. The retrieved CO 2 columns were underestimated by 10 ppm using the HITRAN 2004 data, and improvements resulting from the use of the improved HITRAN database were more pronounced at a higher spectral resolution.

Verification and Correction of Cloud Base and Top Height Retrievals from Ka–band Cloud Radar in Boseong,Korea

In this study,cloud base height（CBH） and cloud top height（CTH） observed by the Ka-band（33.44 GHz） cloud radar at the Boseong National Center for Intensive Observation of Severe Weather during fall 2013（September-November） were verified and corrected.For comparative verification,CBH and CTH were obtained using a ceilometer（CL51） and the Communication,Ocean and Meteorological Satellite（COMS）.During rainfall,the CBH and CTH observed by the cloud radar were lower than observed by the ceilometer and COMS because of signal attenuation due to raindrops,and this difference increased with rainfall intensity.During dry periods,however,the CBH and CTH observed by the cloud radar,ceilometer,and COMS were similar.Thin and low-density clouds were observed more effectively by the cloud radar compared with the ceilometer and COMS.In cases of rainfall or missing cloud radar data,the ceilometer and COMS data were proven effective in correcting or compensating the cloud radar data.These corrected cloud data were used to classify cloud types,which revealed that low clouds occurred most frequently.

EFFECTS OF A CLOUD FILTERING METHOD FOR FENGYUN-3C MICROWAVE HUMIDITY AND TEMPERATURE SOUNDER MEASUREMENTS OVER OCEAN ON RETRIEVALS OF TEMPERATURE AND HUMIDITY

For Microwave Humidity and Temperature sounder(MWHTS) measurements over the ocean, a cloud filtering method is presented to filter out cloud-and precipitation-affected observations by analyzing the sensitivity of the simulated brightness temperatures of MWHTS to cloud liquid water, and using the root mean square error(RMSE)between observation and simulation in clear sky as a reference standard. The atmospheric temperature and humidity profiles are retrieved using MWHTS measurements with and without filtering by multiple linear regression(MLR),artificial neural networks(ANN) and one-dimensional variational(1DVAR) retrieval methods, respectively, and the effects of the filtering method on the retrieval accuracies are analyzed. The numerical results show that the filtering method can improve the retrieval accuracies of the MLR and the 1DVAR retrieval methods, but have little influence on that of the ANN. In addition, the dependencies of the retrieval methods upon the testing samples of brightness temperature are studied, and the results show that the 1DVAR retrieval method has great stability due to that the testing samples have great impact on the retrieval accuracies of the MLR and the ANN, but have little impact on that of the 1DVAR.

An Improved Method for Doppler Wind and Thermo dynamic Retrievals

A variational method is developed to retrieve winds in the first step and then thermodynamic fields in the second step from Doppler radar observations. In the first step, wind fields are retrieved at two time levels: the beginning and ending times of the data assimilation period, simultaneously from two successive volume scans by using the weak form constraints provided by the mass continuity and vorticity equations. As the retrieved wind fields are expressed by Legendre polynomial expansions at the beginning and ending times, the time tendency term in the vorticity equation can be conveniently formulated, and the retrieved winds can be compared with the radar observed radial winds in the cost function at the precise time and position of each radar beam. In the second step, the perturbation pressure and temperature fields at the middle time are then derived from the retrieved wind fields and the velocity time tendency by using the weak form constraints provided by the three momentum equations. The merits of the new method are demonstrated by numerical experiments with simulated radar observations and compared with the traditional least squares methods which consider neither the precise observation times and positions nor the velocity time tendency. The new method is also applied to real radar data for a heavy rainfall event during the 2001 Meiyu season in China.

The Preliminary Discussion of the Potential of GNSS-IR Technology for Terrain Retrievals

The expansion of research and applications of Global Navigation Satellite Systems(GNSS)has revealed the information of reflecting surface in inherent multipath errors.GNSS signals,usually used to measure position,have been demonstrated that they can be used to retrieve water properties including water level,soil moisture,snow depth,and vegetation water content,which are important for climate analysis and water resources monitoring.Reflected GNSS signals with different azimuths can carry information of the corresponding reflecting zone,which means every reflected signal has distinct"signal-tonoise ratio(SNR)characteristics"influenced by specific reflecting zones—and the parameter named"Reflector Height(RH)"deduced from SNR frequency is focused on in this study.Thus,after interpolation of a series of reflector height by coordinates of the footprint,products describing highly detailed terrain over a reflecting footprint can be produced.Data of three GNSS sites in Earth Scope Plate Boundary Observatory,named P025,P351 and P101,was used to evaluate the terrain after calculating the terrain slopes and correcting the footprint following the slopes.A comparison of the results with a digital elevation model(DEM)showed that it is possible to retrieve terrain by GNSS-Interferometric Reflectometry(GNSS-IR);and the comparison with terrain slopes from DEMs in previous research also validated its potential.

The Study of In-Orbit Calibration Accuracy of NOAA Satellite Infrared Sounder and Its Effect on Temperature Profile Retrievals

The calibration accuracy of High Resolution Infrared Radiation Sounder Mod. 2 (HIRS / 2) on NOAA-10 satellite is analyzed in this paper. The non-linear effect in the linear calibration curve induces a deviation of 1.5 degrees (k) of brightness temperature in the tenth channel (8.3 um, water vapor absorption) of the HIRS/2 and the non-linear effect affects the other channels to a different extent. Based on analyzing non- linearity in two-point calibration curve, a tri-point calibration equation is given. A numerical test of effects of the linear and non-linear calibration models on the accuracy of atmospheric temperature retrievals is carried out.

Retrievals of Rain-Rate over Oceans from SSM/IData Using SOM Model

A Self-Organizing Feature Mapping (SOM) network model was developed for the retrievals of rain-rate (RR) over oceans from SSM / I measurement, by using the SSM / I and corresponding radar-derived rain-rate data provided by National Space Development Agency of Japan (NASDA). Based on the frequency-distribution of rain-rate samples, the SOM model was constructed in different rain-rate ranges. The model was first trained by five-sixths of the data, and the other data were used to test the retrieval ability of the model. The retrieval results of the SOM model were compared with two statistically-based algorithms. It is shown that the SOM model provides better retrievals of rain-rate.

IMPROVED ANALYSES AND FORECASTS WITH AIRS TEMPERATURE RETRIEVALS USING THE LOCAL ENSEMBLE TRANSFORM KALMAN FILTER

In this paper we investigate the impact of the Atmospheric Infra-Red Sounder (AIRS) temperature retrievals on data assimilation and the resulting forecasts using the four-dimensional Local Ensemble Transform Kalman Filter (LETKF) data assimilation scheme and a reduced resolution version of the NCEP Global Forecast System (GFS).Our results indicate that the AIRS temperature retrievals have a significant and consistent positive impact in the Southern Hemispheric extratropics on both analyses and forecasts,which is found not only in the temperature field but also in other variables.In tropics and the Northern Hemispheric extratropics these impacts are smaller,but are still generally positive or neutral.

Image Retrieval Based on Vision Transformer and Masked Learning

Deep convolutional neural networks(DCNNs)are widely used in content-based image retrieval(CBIR)because of the advantages in image feature extraction.However,the training of deep neural networks requires a large number of labeled data,which limits the application.Self-supervised learning is a more general approach in unlabeled scenarios.A method of fine-tuning feature extraction networks based on masked learning is proposed.Masked autoencoders(MAE)are used in the fine-tune vision transformer(ViT)model.In addition,the scheme of extracting image descriptors is discussed.The encoder of the MAE uses the ViT to extract global features and performs self-supervised fine-tuning by reconstructing masked area pixels.The method works well on category-level image retrieval datasets with marked improvements in instance-level datasets.For the instance-level datasets Oxford5k and Paris6k,the retrieval accuracy of the base model is improved by 7%and 17%compared to that of the original model,respectively.

Toward Fine-grained Image Retrieval with Adaptive Deep Learning for Cultural Heritage Image

Fine-grained image classification is a challenging research topic because of the high degree of similarity among categories and the high degree of dissimilarity for a specific category caused by different poses and scales.A cul-tural heritage image is one of thefine-grained images because each image has the same similarity in most cases.Using the classification technique,distinguishing cultural heritage architecture may be difficult.This study proposes a cultural heri-tage content retrieval method using adaptive deep learning forfine-grained image retrieval.The key contribution of this research was the creation of a retrieval mod-el that could handle incremental streams of new categories while maintaining its past performance in old categories and not losing the old categorization of a cul-tural heritage image.The goal of the proposed method is to perform a retrieval task for classes.Incremental learning for new classes was conducted to reduce the re-training process.In this step,the original class is not necessary for re-train-ing which we call an adaptive deep learning technique.Cultural heritage in the case of Thai archaeological site architecture was retrieved through machine learn-ing and image processing.We analyze the experimental results of incremental learning forfine-grained images with images of Thai archaeological site architec-ture from world heritage provinces in Thailand,which have a similar architecture.Using afine-grained image retrieval technique for this group of cultural heritage images in a database can solve the problem of a high degree of similarity among categories and a high degree of dissimilarity for a specific category.The proposed method for retrieving the correct image from a database can deliver an average accuracy of 85 percent.Adaptive deep learning forfine-grained image retrieval was used to retrieve cultural heritage content,and it outperformed state-of-the-art methods infine-grained image retrieval.

Image Retrieval with Text Manipulation by Local Feature Modification

The demand for image retrieval with text manipulation exists in many fields, such as e-commerce and Internet search. Deep metric learning methods are used by most researchers to calculate the similarity between the query and the candidate image by fusing the global feature of the query image and the text feature. However, the text usually corresponds to the local feature of the query image rather than the global feature. Therefore, in this paper, we propose a framework of image retrieval with text manipulation by local feature modification(LFM-IR) which can focus on the related image regions and attributes and perform modification. A spatial attention module and a channel attention module are designed to realize the semantic mapping between image and text. We achieve excellent performance on three benchmark datasets, namely Color-Shape-Size(CSS), Massachusetts Institute of Technology(MIT) States and Fashion200K(+8.3%, +0.7% and +4.6% in R@1).

Triplet Label Based Image Retrieval Using Deep Learning in Large Database

Recent days,Image retrieval has become a tedious process as the image database has grown very larger.The introduction of Machine Learning(ML)and Deep Learning(DL)made this process more comfortable.In these,the pair-wise label similarity is used tofind the matching images from the database.But this method lacks of limited propose code and weak execution of misclassified images.In order to get-rid of the above problem,a novel triplet based label that incorporates context-spatial similarity measure is proposed.A Point Attention Based Triplet Network(PABTN)is introduced to study propose code that gives maximum discriminative ability.To improve the performance of ranking,a corre-lating resolutions for the classification,triplet labels based onfindings,a spatial-attention mechanism and Region Of Interest(ROI)and small trial information loss containing a new triplet cross-entropy loss are used.From the experimental results,it is shown that the proposed technique exhibits better results in terms of mean Reciprocal Rank(mRR)and mean Average Precision(mAP)in the CIFAR-10 and NUS-WIPE datasets.

Learning Noise-Assisted Robust Image Features for Fine-Grained Image Retrieval

Fine-grained image search is one of the most challenging tasks in computer vision that aims to retrieve similar images at the fine-grained level for a given query image.The key objective is to learn discriminative fine-grained features by training deep models such that similar images are clustered,and dissimilar images are separated in the low embedding space.Previous works primarily focused on defining local structure loss functions like triplet loss,pairwise loss,etc.However,training via these approaches takes a long training time,and they have poor accuracy.Additionally,representations learned through it tend to tighten up in the embedded space and lose generalizability to unseen classes.This paper proposes a noise-assisted representation learning method for fine-grained image retrieval to mitigate these issues.In the proposed work,class manifold learning is performed in which positive pairs are created with noise insertion operation instead of tightening class clusters.And other instances are treated as negatives within the same cluster.Then a loss function is defined to penalize when the distance between instances of the same class becomes too small relative to the noise pair in that class in embedded space.The proposed approach is validated on CARS-196 and CUB-200 datasets and achieved better retrieval results(85.38%recall@1 for CARS-196%and 70.13%recall@1 for CUB-200)compared to other existing methods.

OSAP‐Loss:Efficient optimization of average precision via involving samples after positive ones towards remote sensing image retrieval

In existing remote sensing image retrieval(RSIR)datasets,the number of images among different classes varies dramatically,which leads to a severe class imbalance problem.Some studies propose to train the model with the ranking‐based metric(e.g.,average precision[AP]),because AP is robust to class imbalance.However,current AP‐based methods overlook an important issue:only optimising samples ranking before each positive sample,which is limited by the definition of AP and is prone to local optimum.To achieve global optimisation of AP,a novel method,namely Optimising Samples after positive ones&AP loss(OSAP‐Loss)is proposed in this study.Specifically,a novel superior ranking function is designed to make the AP loss differentiable while providing a tighter upper bound.Then,a novel loss called Optimising Samples after Positive ones(OSP)loss is proposed to involve all positive and negative samples ranking after each positive one and to provide a more flexible optimisation strategy for each sample.Finally,a graphics processing unit memory‐free mechanism is developed to thoroughly address the non‐decomposability of AP optimisation.Extensive experimental results on RSIR as well as conventional image retrieval datasets show the superiority and competitive performance of OSAP‐Loss compared to the state‐of‐the‐art.

TECMH:Transformer-Based Cross-Modal Hashing For Fine-Grained Image-Text Retrieval

In recent years,cross-modal hash retrieval has become a popular research field because of its advantages of high efficiency and low storage.Cross-modal retrieval technology can be applied to search engines,crossmodalmedical processing,etc.The existing main method is to use amulti-label matching paradigm to finish the retrieval tasks.However,such methods do not use fine-grained information in the multi-modal data,which may lead to suboptimal results.To avoid cross-modal matching turning into label matching,this paper proposes an end-to-end fine-grained cross-modal hash retrieval method,which can focus more on the fine-grained semantic information of multi-modal data.First,the method refines the image features and no longer uses multiple labels to represent text features but uses BERT for processing.Second,this method uses the inference capabilities of the transformer encoder to generate global fine-grained features.Finally,in order to better judge the effect of the fine-grained model,this paper uses the datasets in the image text matching field instead of the traditional label-matching datasets.This article experiment on Microsoft COCO(MS-COCO)and Flickr30K datasets and compare it with the previous classicalmethods.The experimental results show that this method can obtain more advanced results in the cross-modal hash retrieval field.