Model For Correction Research Articles

Global Gridded Aerosol Models Established for Atmospheric Correction Over Inland and Nearshore Coastal Waters

AbstractThe use of standard aerosol models for atmospheric correction over inland and nearshore coastal waters generates large uncertainties due to their inability to appropriately characterize aerosols in those regions. To address this problem, we developed new aerosol models using data acquired from 1,475 Aerosol Robotic Network (AERONET) sites distributed in inland and coastal regions throughout the globe. We established spatially gridded (5° × 5°) aerosol models on a climatological monthly scale, with a total of 3,207 new aerosol models developed for 310 grid cells worldwide. The monthly aerosol model for each grid was derived using mean monthly values of microphysical aerosol properties from AERONET data. The optical properties (i.e., single scattering albedo and Ångström coefficient) of our global gridded aerosol models demonstrated considerable spatial and seasonal dynamics and deviated substantially from the aerosol models currently utilized by the National Aeronautics and Space Administration (NASA). We built associated lookup tables for all aerosol models and incorporated them into two atmospheric correction algorithms: the combined near‐infrared and shortwave infrared algorithm (NIR‐SWIR) and the Atmospheric Correction algorithm for inLand and Nearshore Coastal waters (ACLANC). For both algorithms, the accuracies of atmospherically corrected satellite remote sensing reflectance (Rrs) improved markedly when the new aerosol models were used, and improvements were noted across all spectral bands and across all six inland and nearshore water bodies used in this study. Our new aerosol models can be easily applied to all ocean color satellite missions to obtain accurate Rrs products for inland and coastal waters worldwide.

Spatiotemporal Assessment and Correction of Gridded Precipitation Products in North Western Morocco

Accurate and spatially distributed precipitation data are fundamental to effective water resource management. In Morocco, as in other arid and semi-arid regions, precipitation exhibits significant spatial and temporal variability. Indeed, there is an intra- and inter-annual variability and the northwest is rainier than the rest of the country. In the Bouregreg watershed, this irregularity, along with a sparse gauge network, poses a major challenge for water resource management. In this context, remote sensing data could provide a viable alternative. This study aims precisely to evaluate the performance of four gridded daily precipitation products: three IMERG-V06 datasets (GPM-F, GPM-L, and GPM-E) and a reanalysis product (ERA5). The evaluation is conducted using 11 rain gauge stations over a 20-year period (2000–2020) on various temporal scales (daily, monthly, seasonal, and annual) using a pixel-to-point approach, employing different classification and regression metrics of machine learning. According to the findings, the GPM products showed high accuracy with a low margin of error in terms of bias, RMSE, and MAE. However, it was observed that ERA5 outperformed the GPM products in identifying spatial precipitation patterns and demonstrated a stronger correlation. The evaluation results also showed that the gridded precipitation products performed better during the summer months for seasonal assessment, with relatively lower accuracy and higher biases during rainy months. Furthermore, these gridded products showed excellent performance in capturing different precipitation intensities, with the highest accuracy observed for light rain. This is particularly important for arid and semi-arid regions where most precipitation falls under the low-intensity category. Although gridded precipitation estimates provide global coverage at high spatiotemporal resolutions, their accuracy is currently insufficient and would require improvement. To address this, we employed an artificial neural network (ANN) model for bias correction and enhancing raw precipitation estimates from the GPM-F product. The results indicated a slight increase in the correlation coefficient and a significant reduction in biases, RMSE, and MAE. Consequently, this research currently supports the applicability of GPM-F data in North Western Morocco.

Analog sun sensor measurement correction using deep neural network

Sun sensors are commonly used attitude sensors because of their low cost, mass, volume, and power consumption. Analog Sun sensors (ANSS), which are smaller but usually less accurate than digital ones, are especially preferred for small satellite missions. One of the main reasons for the lesser accuracy of analog Sun sensors is being prone to external errors, most prominently the Earth's albedo. This study proposes an analog Sun sensor calibration method using the Deep Neural Network (DNN). The main contribution of the proposed algorithm is that it does not require any model for measurement correction. The method is tested with simulations and real data from an Earth-imaging spacecraft. Results show that the error in the Sun direction measurements, which can be as high as 10°, can be decreased to a level of 0.5° by using the DNN for calibration. Moreover, testing in different scenarios verifies that the DNN can correct the measurements for periods as long as 7 days without requiring excessive training periods, even when the spacecraft is not in the same flight configuration for which the DNN was trained.

Models of interphase drag force from direct numerical simulations of upward turbulent particle-laden channel flows

Data from the interfaced-resolved direct numerical simulations (IR-DNSs) of upward turbulent channel flow laden with finite-size heavy particles at particle concentration ranging from 0.84% to 2.36% are used to develop improved drag correlations which account for the effect of turbulence on drag enhancement. First, the contribution of drag nonlinearity to drag enhancement for the turbulent flow over a fixed sphere is analyzed with the consideration of velocity fluctuation anisotropy. The nonlinear drag enhancement is generally smaller than the actual drag enhancement in our numerical simulations. Second, several models for turbulence correction are presented with different choice of model parameters, and it is found that the turbulent intensity normalized by the mean slip velocity and the particle size normalized by the Kolmogorov scale are most important for modeling the effect of turbulence on particle drag, while the anisotropy of fluctuating velocity and the particle Reynolds number are less important.

ADAPTIVE IMAGE ENHANCEMENT MODEL FOR THE ROBOT VISION SYSTEM

Robotics is one of the important trends in the current development of science and technology. Most modern robots and drones have their own vision system, including a video camera, which they use to take digital photos and video streams. These data are used to analyze the situation in the robot's camera field of view, as well as to determine a real-time robot's behavior algorithm. In this regard, the novelty of the paper is special polynomial mathematical model and method for adaptive gradational correction of a digital image. The proposed model and method make it possible to independently adjust to brightness scales and image formats and optimally perform gradational image correction in various lighting conditions. Thus, ensuring the efficiency of the entire subsequent cycle of image analysis in the robot's vision system. In addition, the paper presents the results of numerous experiments of such gradational correction for images of various classes, as well as conditions of reduced and increased levels of illumination of the field of view objects. Conclusions and recommendations are given regarding the practical application of the proposed model and method.

Rapid Correction of Turbidity and CDOM Interference on Three-Dimensional Fluorescence Spectra of Live Algae Based on Deep Learning

In natural aquatic environments, the existence of colored dissolved organic matter (CDOM), suspended particles, and colloids can cause scattering and reflection of light and even emit fluorescence itself. Such interference negatively impacts algal fluorescence, further making it unreliable to measure the algal concentration using three-dimensional excitation–emission matrix (3D-EEM) fluorescence spectroscopy. In this study, we proposed a novel algal fluorescence anti-interference network (AFAI-Net) based on a convolutional neural network. The main procedure of this model can be divided into two parts: (1) to quickly determine if there is an interference of CDOM or turbidity in the detected algal samples; (2) to correct the interfered samples and output the fluorescent components of the algae. We trained the model using the 3D-EEMs of pure algal samples (non-interfered) and mixed samples of algae and CDOM or turbidity (interfered); as a result, the well-trained model achieved a total classification accuracy of 96.82%, and the RMSE of CDOM and turbidity removal fitting effects were 0.2274 and 0.3423, respectively. Compared with the non-negative weighted least squares (NNLS) regression analysis method, using the CNN model for CDOM correction resulted in 13.11%, 0.65%, and 5.69% reductions in the average deviation rate for PD, PG, and CM, respectively. Furthermore, the spectra corrected by the model predicted algal densities that were closer to the true algal densities. This study provides a new way to remove non-algal factors that affect algal fluorescence spectra in water bodies, which is beneficial to monitoring eutrophication and red tide in aquatic systems.

The mantle transition zone beneath eastern North America: Receiver functions and tomographic velocity models

The eastern continental margin of North America, despite being a passive margin at present, records a comprehensive tectonic history of both mountain building and rifting events. This record is punctuated by several igneous events, including those associated with the Great Meteor and Bermuda hotspots. To gain a better understanding of the state of the mantle beneath this region, we employ the massive quantity of seismic data recorded by the USArray to image the mantle transition zone beneath eastern North America. To construct these images, we first calculate P-to-s receiver functions using an iterative time-domain deconvolution algorithm. These receiver functions are then automatically filtered by their quality, using a set of rigorous criteria, and subsequently summed using common conversion point stacking. We present several cross sections through these stacks, which show remarkable features such as a thinned transition zone beneath the independently observed northern Appalachian and central Appalachian low-wavespeed anomalies, as well as a thickened transition zone beneath western Tennessee associated with the Laramide slab stagnating at depth. In addition to discussing these geologically relevant features, we perform a technical analysis of the effects of using various seismic velocity models for the moveout correction of our receiver functions. We find that the thickness of the mantle transition zone under eastern North America is a robust measurement, while the resolved depths of the 410 and 660 km discontinuities are model dependent.

Study on optimization model control method of light and temperature coordination of greenhouse crops with benefit priority

Facility agriculture is the foundation of modern agriculture. In recent years, energy consumption has become a major bottleneck restricting the development of greenhouse. However, with the support of intelligent control methods, the greenhouse energy-saving market with cost reduction and efficiency has gained increasing attention. Therefore, there is a great potential to expand the research on coordinated energy conservation and optimal control of greenhouse environment, with the objective of improving greenhouse yield and reducing production costs. In this study, a phototemperature coupled net photosynthetic rate (Pn) model was established by using the rectangular hyperbolic correction model of temperature correction, and the model parameters were obtained by Nlinfit fitting. Based on the NSGA-II algorithm, the Pareto optimal frontier was obtained with the objective of minimizing energy consumption and maximizing predicted crop yield. The artificial fish swarm algorithm was used to find the maximum point for the three-dimensional interpolation network of the established crop model. Different evaluation strategies were used to evaluate the optimal solution. The results showed that the average power consumption before and after optimization was reduced by 4.31 %, when Pn remained basically unchanged. When the power consumption remained basically unchanged, the average increase in Pn was about 9.59 %.

Unraveling-paired dynamical maps recover the input of quantum channels

We explore algebraic and dynamical consequences of unraveling general time-local master equations. We show that the ‘influence martingale’, the paramount ingredient of a recently discovered unraveling framework, pairs any time-local master equation with a one parameter family of Lindblad–Gorini–Kossakowski–Sudarshan master equations. At any instant of time, the variance of the influence martingale provides an upper bound on the Hilbert–Schmidt distance between solutions of paired master equations. Finding the lowest upper bound on the variance of the influence martingale yields an explicit criterion of ‘optimal pairing’. The criterion independently retrieves the measure of isotropic noise necessary for the structural physical approximation of the flow the time-local master equation with a completely positive flow. The optimal pairing also allows us to invoke a general result on linear maps on operators (the ‘commutant representation’) to embed the flow of a general master equation in the off-diagonal corner of a completely positive semi-group which in turn solves a time-local master equation that we explicitly determine. We use the embedding to reverse a completely positive evolution, a quantum channel, to its initial condition thereby providing a protocol to preserve quantum memory against decoherence. We thus arrive at a model of continuous-time error correction by a quantum channel.

Adaptive model for the water depth bias correction of bathymetric LiDAR point cloud data

The water depth bias of LiDAR point cloud data has to be corrected. The previous models, which used the fixed parameters for an entire water area, cannot efficiently correct the bias due to the different water environments. Therefore, this paper develops an adaptive model for the correction of the water depth bias. A coordinate system, in which the water depth is taken as the X-axis and the water depth bias is taken as the Y-axis, is defined. All of the sample points are normalized, and then projected into the defined coordinate system according to their water depths and water depth biases. Second, the scatter points are clustered into several different clusters using the developed subdivision algorithm. With the clusters, an entire water area is subdivided into several sub-regions. Finally, each sub-region is fitted using a model, which is used to correct the water depth bias. Experimental verification and comparison analysis are conducted in three different environments: an indoor tank, an outdoor pond and the Beihai sea. The experimental results demonstrate that the Mean Absolute Error (MAE) and the Root Mean Squared Error (RMSE) from the adaptive model are reduced by approximately 61% and 59%, respectively, relative to those from the traditional models. Therefore, it can be concluded that the proposed model can adapt to changes of the different water environments and achieve a higher accuracy for water depth bias correction than traditional methods do.

Study on a benchmark finite element model of a cable-stayed bridge based on ABAQUS and IM-HPO

To solve the problem that finite element analysis and calculation need to be repeated during model updates, which leads to tedious and inefficient correction work, ABAQUS (a powerful finite element software for engineering simulation) is used to develop a user-friendly graphical user interface platform for automatic completion of this update for cable-stayed bridges. In this paper, a method for establishing the benchmark finite element model of cable-stayed bridges based on ABAQUS and influence matrix (IM)-hunter prey optimization (HPO) is proposed. The method is based on the secondary development of ABAQUS to realize the modules of sensitivity analysis, influence matrix, and optimization. According to the sensitivity analysis module, the influence of the change in the finite element model parameters on the static response of a cable-stayed bridge is studied, and the cable area is used as the parameter to be corrected. The influence matrix module is used to obtain the IM of the cable area to girder displacement and cable force under a static load. The objective function is constructed by combining the measured displacement and cable force data using the optimization module, and the global optimal solution of the parameters to be corrected in the objective function is found by HPO. Finally, the solution is sent to the finite element model for correction, and the corrected displacement and cable force data are compared with the measured data. The results show that the modified calculation results obtained by this method are in good agreement with the measured results. The modified finite element model can be used as presented.

Multiple step saccades are generated by internal real-time saccadic error correction.

Multiple step saccades (MSSs) are an atypical form of saccade that consists of a series of small-amplitude saccades. It has been argued that the mechanism for generating MSS is due to the automatic saccadic plan. This argument was based on the observation that trials with MSS had shorter saccadic latency than trials without MSS in the reactive saccades. However, the validity of this argument has never been verified by other saccadic tasks. Alternatively, we and other researchers have speculated that the function of MSS is the same as that of the corrective saccade (CS), i.e., to correct saccadic errors. Thus, we propose that the function of the MSS is also to rectify saccadic errors and generated by forward internal models. The objective of the present study is to examine whether the automatic theory is universally applicable for the generation of MSSs in various saccadic tasks and to seek other possible mechanisms, such as error correction by forward internal models. Fifty young healthy subjects (YHSs) and fifty elderly healthy subjects (EHSs) were recruited in the present study. The task paradigms were prosaccade (PS), anti-saccade (AS) and memory-guided saccade (MGS) tasks. Saccadic latency in trials with MSS was shorter than without MSS in the PS task but similar in the AS and MGS tasks. The intersaccadic intervals (ISI) were similar among the three tasks in both YHSs and EHSs. Our results indicate that the automatic theory is not a universal mechanism. Instead, the forward internal model for saccadic error correction might be an important mechanism.

TECHNOLOGY FOR GRAMMATICAL ERRORS CORRECTION IN UKRAINIAN TEXT CONTENT BASED ON MACHINE LEARNING METHODS

Context. Most research in grammatical and stylistic error correction focuses on error correction in English-language textual content. Thanks to the availability of large data sets, a significant increase in the accuracy of English grammar correction has been achieved. Unfortunately, there are few studies on other languages. Systems for the English language are constantly developing and currently actively use machine learning methods: classification (sequence tagging) and machine translation. A large amount of parallel or manually labelled data is required to build a high-quality machine learning model for correcting grammatical/stylistic errors in the texts of those morphologically complex languages. Manual data annotation requires a lot of effort by professional linguists, which makes the creation of text corpora, especially in morphologically rich languages, mainly Ukrainian, a time- and resource-consuming process.
 Objective of the study is to develop a technology for correcting errors in Ukrainian-language texts based on machine learning methods using a small set of annotated parallel data.
 Method. For this study, machine learning algorithms were selected when developing a system for correcting errors in Ukrainianlanguage texts using an optimal pipeline, including pre-processing and selecting text content and generating features in small annotated data corpora. The neural network’s use with a new architecture, a review of state-of-the-art methods, and a comparison of different pipeline stages will make it possible to determine such a combination of them, allowing a high-quality error correction model in Ukrainian-language texts.
 Results. A machine learning model for error correction in Ukrainian-language texts has been developed. A universal scheme for creating an error correction system for different languages is proposed. According to the results, the neural network can correct simple sentences written in Ukrainian. However, creating a full-fledged system will require spell-checking using dictionaries and checking rules, both simple and based on the result of parsing dependencies or other features. The pre-trained neural translation model mT5 has the best performance among the three models. To save computing resources, it is also possible to use a pre-trained BERT-type neural network as an encoder and a decoder. Such a neural network has half the number of parameters as other pretrained machine translation models and shows satisfactory results in correcting grammatical and stylistic errors.
 Conclusions. The created model shows excellent classification results on test data. The calculated machine translation quality metrics allow only a partial comparison of the models since most of the words and phrases in the original and corrected sentences are the same. The best value for both BLEU (0.908) and METEOR (0.956) is obtained for mT5, which is consistent with the case study in which the most accurate error corrections without changing the initial value of the sentence are obtained for such a neural network. The M2M100 has a higher BLEU score (0.847) than the “Ukrainian Roberta” Encoder-Decoder (0.697). However, subjectively evaluating the results of the correction of examples, the M2M100 does a much worse job than the other two models. For METEOR, M2M100 (0.925) also has a higher score than the “Ukrainian Roberta” Encoder-Decoder (0.876).

Training model for the fetal myelomeningocele correction with multiportal endoscopic technique.

The recent history of myelomeningocele has shown that treatment during the fetal life may reduce the risk of developing hydrocephalus in individuals by approximately 50%. Thus, a significant advancement involves fetal surgery performed through an endoscopic technique in which portals are placed to introduce the forceps and laparoscopic instruments. However, the development of this technique requires training; therefore, this study aimed to develop a training model for fetal myelomeningocele repair technique with multi-portal endoscopy. Two stages of endoscopic technique development were performed. The first stage consisted of exercises in order to familiarize the surgeon with 2D-vision endoscopic surgery, associated with the application of exercises focused on surgical skills, such as the development of laparoscopic knots in a synthetic model. The second stage involved the creation and application of the stages of myelomeningocele closure with a non-living animal model consisting of a chicken breast to simulate the myelomeningocele and a basketball to simulate the gravid uterus, in which perforations were made to introduce vascular introducers (portals) that, as in vivo, are used as portals (trocars) for the introduction of laparoscopic instruments. Overall, two different scenarios with three portals and two portals were tested. In three-portal simulator, the triangular apex trocar was used for the introduction of 4-mm 0° or 30° optics or even Minop type neurodoscope (Aesculap®, Germany) that was operated by the assistant surgeon; the other two portals are used for the introduction of laparoscopic instruments. Thus, the surgeon is able to perform maneuvers bimanually since dissection to laparoscopic sutures. In two-portal simulator, the surgeon and assistant stay side by side and one of the portals is used for the optic and the other for the laparoscopic instruments. There is no possibility of bimanual dissection in this method. Realistic simulation models for endoscopic fetal surgery for myelomeningocele correction are easily performed and help develop the necessary skills for fetal surgery teams.

Surgical Simulation of Posterior Sagittal Anorectoplasty for Rectovestibular Fistula: Low-Cost High-Fidelity Animal-Tissue Model

PurposeTo provide a high-fidelity, animal tissue-based model for the advanced surgical simulation of a Posterior Sagittal Anorectoplasty (PSARP) for rectovestibular fistula in anorectal malformation (ARM). Materials and methodsA chicken cadaver was used to assess the feasibility of simulating a PSARP for rectovestibular fistula in ARM. No modification was required to implement the surgical simulation. ResultsA detailed description of the high-fidelity surgical simulation model is provided. The PSARP can be simulated while providing realistic anatomy (e.g. common wall between rectovestibular fistula and vagina), adequate rectal size, location and placement of the rectovestibular fistula, and proximity to the vagina. Haptic conditions of the tissue resemble human tissue and operative conditions as well. DiscussionConcerning the decreased exposure of index cases of pediatric surgical trainees and pediatric surgeons in practice, simulation-based training can provide means to acquire or maintain the necessary skills to perform complex surgical procedures [1–5]Surgical simulation models for ARM are limited. Few low-cost trainers are available with predominant artificial and mostly unrealistic tissue [6–8] Animal models have the advantage of realistic multilayer tissue haptic feedback [6]. ConclusionWe provide a low-cost, high-fidelity model for correcting a rectovestibular fistula in a child with ARM, a complex operative procedure with low incidence but high-stake outcomes. The described tissue model utilizing the chicken cloaca anatomy provides a high-fidelity model for operative correction of rectovestibular ARM. For simulation purposes in the treatment of ARM, this model appears to be promising in terms of providing realistic pathology and haptic feedback in pediatric dimensions. Level of evidenceV.

The X chromosome still has a lot to reveal - revisiting Hermann Henking's work on firebugs.

In 1891, the existence of an X chromosome was noted for the first time. Hermann Henking was studying spermatocyte divisions of the firebug Pyrrhocoris apterus and observed that one chromosome behaved differently than all of the rest of the chromosomes. Henking called this chromosome 'Element x'. Henking's discovery of the X element (later called X chromosome) initiated more than a century of fascinating genetics and cell biology, forming the foundation of several avenues of research in biology. His work led to exploration of a number of questions in a wide range of model systems and very soon to the abandonment of the firebug as a model for studies on the behavior of chromosomes in meiosis. Here, we argue that studies on both bivalent and univalent chromosome behavior in general, and work on how to solve chromosome lagging to prevent aneuploidy in particular, should lead us back to using the firebug as a model for error correction during cell division.

Financial stability in the Indonesian monetary policy analysis

Financial stability is one of the main factors for a country’s economic sustainability nowadays. Financial instability has adverse effects on the economy which can lead to the financial crisis. This research tries to answer how monetary policy related to Indonesian financial stability in the short and long term, how specific the relations between its parts such as money supply, interest rate, and exchange rate to indonesian financial stability, and which one of these parts that have ways more effective to influence Indonesian financial stability. This research data use time series quarterly as quantitative data in Indonesia. By employing econometric models of error correction on quarter data that consist of the stationary stochastic process, long-run model and cointegration approach and error correction model, this model can indicate how quickly the effect of the money supply, interest rate or exchange rate is fully accepted by the financial stability in the short term. The finding of this study showed that the increase in interest rate began to be able to improve indonesian financial stability responsively in four times the period, namely an increase of one percent in interest rate could increase 0.4161 percent of financial stability in Indonesia.

Development and characterization of a passive, bio-inspired flow-tracking sensor

The effective natural transport of seeds in turbulent atmospheric flows is found across a myriad of shapes and sizes. However, to develop a sensitive passive sensor required for large-scale (in situ) flow tracking measurements, systems suffer from inertial lag due to the increased size and mass needed for optical visibility, or by carrying a sensor payload, such as an inertial measurement unit (IMU). While IMU-based flow sensing is promising for beyond visual line-of-sight applications, the size and mass of the sensor platform results in reduced flow fidelity and, hence, measurement error. Thus, to extract otherwise inaccessible flow information, a flow-physics-based tracer correction is developed through the application of a low-order unsteady aerodynamic model, inspired by the added-mass concept. The technique is evaluated using a sensor equipped with an IMU and magnetometer. A spherical sensor platform, selected for its symmetric geometry, was subject to two canonical test cases including an axial gust as well as the vortex shedding generated behind a cylinder. Using the measured sensor velocity and acceleration as inputs, an energized-mass-based dynamic model is used to back-calculate the instantaneous flow velocity from the sensor measurements. The sensor is also tracked optically via a high-speed camera while collecting the inertial data onboard. For the 1D test case (axial gust), the true (local) wind speed was estimated from the energized-mass-based model and validated against particle image velocimetry measurements, exhibiting good agreement with a maximum error of 10%. For the cylinder wake (second test case), the model-based correction enabled the extraction of the velocity oscillation amplitude and vortex-shedding frequency, which would have otherwise been inaccessible. The results of this study suggest that inertial (i.e. large and heavy) IMU-based flow sensors are viable for the extraction of Lagrangian tracking at large atmospheric scales and within highly-transient (turbulent) environments when coupled with a robust dynamic model for inertial correction.

Validation of a hands-on porcine bladder model for endoscopic correction of vesicoureteral reflux with bulking agents

Validation of a hands-on porcine bladder model for endoscopic correction of vesicoureteral reflux with bulking agents

Gaussian Process and Deep Learning Atmospheric Correction

Atmospheric correction is the processes of converting radiance values measured at a spectral sensor to the reflectance values of the materials in a multispectral or hyperspectral image. This is an important step for detecting or identifying the materials present in the pixel spectra. We present two machine learning models for atmospheric correction trained and tested on 100,000 batches of 40 reflectance spectra converted to radiance using MODTRAN, so the machine learning model learns the radiative transfer physics from MODTRAN. We created a theoretically interpretable Bayesian Gaussian process model and a deep learning autoencoder treating the atmosphere as noise. We compare both methods for estimating gain in the correction model to process for estimating gain within the well-know QUAC method which assumes a constant mean endmember reflectance. Prediction of reflectance using the Gaussian process model outperforms the other methods in terms of both accuracy and reliability.