Relative Root Mean Square Error Research Articles

Image synthesis of effective atomic number images using a deep convolutional neural network-based generative adversarial network.

The effective atomic numbers obtained from dual-energy computed tomography (DECT) can aid in characterization of materials. In this study, an effective atomic number image reconstructed from a DECT image was synthesized using an equivalent single-energy CT image with a deep convolutional neural network (CNN)-based generative adversarial network (GAN). The image synthesis framework to obtain the effective atomic number images from a single-energy CT image at 120 kVp using a CNN-based GAN was developed. The evaluation metrics were the mean absolute error (MAE), relative root mean square error (RMSE), relative mean square error (MSE), structural similarity index (SSIM), peak signal-to-noise ratio (PSNR), and mutual information (MI). The difference between the reference and synthetic effective atomic numbers was within 9.7% in all regions of interest. The averages of MAE, RMSE, MSE, SSIM, PSNR, and MI of the reference and synthesized images in the test data were 0.09, 0.045, 0.0, 0.89, 54.97, and 1.03, respectively. In this study, an image synthesis framework using single-energy CT images was constructed to obtain atomic number images scanned by DECT. This image synthesis framework can aid in material decomposition without extra scans in DECT.

A Methodology for Checking the Validity of an Existing Contour Map Using RTK-GPS and GIS

This paper presents a methodology for checking the validity of the height information of an existing contour map using RTK-GPS measurements and GIS analysis. In particular, 3D profiles were measured in the field by RTK-GPS. The same 3D profiles were extracted from a scanned and geo-referenced contour map using GIS. The height information from the two data sets of 3D profiles were compared in the absolute and relative sense. Thiscomparison reveals two aspects about the two data sets. First, the two data sets from the RTK-GPS measurementsand the existing maps are very similar in terms of shape, which suggest that both of them they capture the same underlying trend of the topography. Second, the maximum relative RMSE is  0.939 m in the first 3D profile that goes from the west to the east. The relative RMSE is a very important measure since it does not depend on the changes that may happened to the vertical datum (benchmark). These two findings leave us with two options. First, to accept the existing contour map; and this acceptance should be judged in light of the overall requirements and the design criteria of the project. For example, a global compensation can be made to accommodate the maximum variation in the topographic surface for an irrigation project. This compensation does not give any information about the local variation of the topography that may impact, for instance, a pivot irrigation design. Second, to redo the survey work. Indeed, both options are associated with obvious and hidden risks that may not be trivial during thedesign stage of the project. The first option can be accepted on the basis of a global compensation parameters that will be obtained from the comparison tests. And the second option will come with an instantaneous financial burden.

Estimating leaf nitrogen content by coupling a nitrogen allocation model with canopy reflectance

The area-based leaf nitrogen content (LNCA) has gained increasing attention recently in ecology and plant physiology, and the mass-based nitrogen content (LNCM) has been extensively used by agronomists for making fertilization recommendations. Remote sensing has been emerging as an important tool to quantify LNCA and LNCM across multiple observation scales, such as leaf and canopy. Most models for estimating LNCA and LNCM are based on their empirical relationships with chlorophyll-related vegetation indices (VIs), which cannot capture the dynamic allocation of nitrogen into different components throughout the growing season. This study proposed a simplified nitrogen allocation model by dividing the total nitrogen into photosynthetic nitrogen (PN) and non-photosynthetic nitrogen with the purpose of estimating LNCA and LNCM in rice and winter wheat. The nitrogen allocation model was calibrated with an independent leaf-scale dataset (R2 = 0.82) and then applied to the canopy scale without the need to adjust the model coefficients. The estimated fractions of PN were 71% and 78% for rice and wheat leaves, respectively, which were close to the reported value (∼75%) for common C3 plants. Overall, the nitrogen allocation-based models were applicable for estimating both LNCA and LNCM with relative root mean square error (RRMSE) values <15% for the jointing and booting stages. In addition, our novel estimation model was generic for rice and wheat, yielding the best estimations for LNCM (R2 = 0.63) across the entire growing season compared with traditional empirical models based on VIs. This study represents a first attempt to couple nitrogen allocation theory with remote sensing data to improve the spectroscopic estimation of LNCA and LNCM.

Double-Slope Solar Still Productivity Based on the Number of Rubber Scraper Motions

In low-latitude areas less than 10° in latitude angle, the solar radiation that goes into the solar still increases as the cover slope approaches the latitude angle. However, the amount of water that is condensed and then falls toward the solar-still basin is also increased in this case. Consequently, the solar yield still is significantly decreased, and the accuracy of the prediction method is affected. This reduction in the yield and the accuracy of the prediction method is inversely proportional to the time in which the condensed water stays on the inner side of the condensing cover without collection because more drops will fall down into the basin of the solar-still. Different numbers of scraper motions per hour (NSM), that is, 1, 2, 3, 4, 5, 6, and 7, are implemented to increase the hourly yield of solar still (HYSS) of the double-slope solar still hybrid with rubber scrapers (DSSSHS) in areas at low latitudes and develop an accurate model for forecasting the HYSS. The proposed model is developed by determining the best values of the constant factors that are associated with NSM, and the optimal values of exponent (n) and the unknown constant (C) for the Nusselt number expression (Nu). These variables are used in formulating the models for estimating HYSS. The particle swarm optimization (PSO) algorithm is used to solve the optimization problem, thereby determining the optimal yields. Water that condensed and accumulated inside the condensing glass cover of the DSSSHS is collected by increasing NSM. This process increases in the specific productivity of DSSSHS and the accuracy of the HYSS prediction model. Results show that the proposed model can consistently and accurately estimate HYSS. Based on the relative root mean square error (RRMSE), the proposed model PSO–HYSS attained a minimum value (2.81), whereas the validation models attained Dunkle’s (78.68) and Kumar and Tiwari’s (141.37).

Simulations of nitrous oxide emissions and global warming potential in a C4 turfgrass system using process-based models

Nitrous oxide (N2O) is an ozone-depleting gas and important greenhouse gas (GHG) that has been implicated in global climate change. Irrigated and N-fertilized turfgrass systems cover large areas of land and emit N2O. DAYCENT and Denitrification-Decomposition (DNDC) are two widely used process-based models that predict GHG fluxes in agricultural lands. Despite many experiments applying these models to other cropping systems, they have not been applied to warm-season (C4) turfgrass systems. Our objectives were to 1) calibrate DAYCENT and DNDC for N2O emissions from ‘Meyer’ zoysiagrass (Zoysia japonica Steud., referred to as zoysia); 2) validate both models and compare their prediction accuracies; and 3) predict long-term impacts of different N fertilization and irrigation management. A combination of global sensitivity analysis and a Bayesian method was used to calibrate DAYCENT and DNDC, followed by validation, using measurements from zoysia field studies. Validation results indicated DAYCENT (R2 = 0.30–0.90; relative RMSE = 30–163%) outperformed DNDC (R2 = 0.01–0.38; relative RMSE = 119–183%) in biweekly N2O fluxes. Annual N2O emissions from DAYCENT were validated within − 54 to + 14% of estimates interpolated from measurements of different N fertilization and irrigation management, whereas DNDC simulations generally underestimated emissions by − 24 to − 85%. DAYCENT, but not DNDC, adequately simulated the impacts of irrigation and N-fertilization practices on N2O emissions in zoysia, a C4 turfgrass. DAYCENT predicted that typical N-fertilization and irrigation practices in fairway zoysia turf would reduce net global warming potential (GWP) better than no N fertilization by encouraging soil carbon sequestration in the first 30–45 years of establishment. Instead of fixed-amount or no N inputs, gradually reducing N fertilization to maintain turfgrass growth over decades would mitigate the increases of net GWP and better alleviate climate challenges in this century.

Development of a filter-based near-surface remote sensing system to retrieve far-red sun-induced chlorophyll fluorescence

Observations of sun-induced chlorophyll fluorescence (SIF) by remote sensing have improved our understanding of the structural and physiological dynamics of vegetation. Substantial efforts have been made to measure SIF with ground-based sensing systems, but field observation data for various plant functional types are still sparse. This is partly due to the limited availability of commercial SIF measurement systems, the relatively high cost of hyperspectral spectroradiometers, and the difficulties of sensor calibration and maintenance in the field. We developed a filter-based smart near-surface remote sensing system for SIF (4S-SIF) to overcome the technical challenges of monitoring SIF in the field, which also decreased the sensor cost, thus enabling more comprehensive spatial sampling. To retrieve SIF, we combined ultra-narrow bandpass filters (full width half maximum <1.3 nm) and photodiode detectors to observe electromagnetic radiation at specific wavelengths (757, 761, and 770 nm). We confirmed that the spectral and radiometric performance of the bandpass filters was satisfactory to retrieve SIF by comparing them to a high-spectral-resolution spectroradiometer that served as a reference. In particular, we confirmed that the digital numbers (DNs) from 4S-SIF exhibited linear relationships with the DN from the reference spectroradiometer in each band (R2 > 0.99). In addition, we developed equations to correct for temperature-induced changes in filter transmittance, such that SIF can be reliably extracted in outdoor environments without the need to actively stabilize the temperature. Furthermore, we confirmed that the SIF signal from 4S-SIF had a strong linear relationship with the reference spectroradiometer-based SIF. Importantly, this relationship held even when the physiological mechanisms of the plant were altered by a herbicide treatment that induced substantial changes in the SIF signal (R2 = 0.85, relative RMSE = 0.22), which indicated that 4S-SIF could be used to retrieve SIF. We believe that 4S-SIF will be a useful tool for collecting in-situ SIF data across multiple spatial and temporal scales.

A Novel End-to-end Network Based on a bidirectional GRU and a Self-Attention Mechanism for Denoising of Electroencephalography Signals

Electroencephalography (EEG) signals are nonlinear and non-stationary sequences that carry much information. However, physiological signals from other body regions may readily interfere with EEG signal capture, having a significant unfavorable influence on subsequent analysis. Therefore, signal denoising is a crucial step in EEG signal processing. This paper proposes a bidirectional gated recurrent unit (GRU) network based on a self-attention mechanism (BG-Attention) for extracting pure EEG signals from noise-contaminated EEG signals. The bidirectional GRU network can simultaneously capture past and future information while processing continuous time sequence. And by paying different levels of attention to the content of varying importance, the model can learn more significant feature of EEG signal sequences, highlighting the contribution of essential samples to denoising. The proposed model is evaluated on the EEGdenoiseNet data set. We compared the proposed model with a fully connected network (FCNN), the one-dimensional residual convolutional neural network (1D-ResCNN), and a recurrent neural network (RNN). The experimental results show that the proposed model can reconstruct a clear EEG waveform with a decent signal-to-noise ratio (SNR) and the relative root mean squared error (RRMSE) value. This study demonstrates the potential of BG-Attention in the pre-processing phase of EEG experiments, which has significant implications for medical technology and brain-computer interface (BCI) applications.

Using multisource satellite products to estimate forest aboveground biomass in Oita prefecture: a novel approach with improved accuracy and computational efficiency

Accurate estimation of forest aboveground biomass (AGB) using satellite information is crucial for quantitatively evaluating forest carbon stock for climate change mitigation. However, with the emergence of an enormous number of identical data, selecting the most valuable dataset for estimation is essential to improve model efficiency and accuracy. Here, we present a novel framework using principal component analysis (PCA)-derived textures to tackle redundant variables and the integration of multisource satellite images using the random forest regression algorithm with recursive feature elimination to map the AGB of deciduous broadleaf forests in Oita prefecture, Japan. Our results highlighted that the model using PCA and derivative features had the best performance [R2 = 0.69, root mean square error (RMSE) = 33.59 Mg/ha, mean absolute error = 25.01 Mg/ha, relative RMSE = 0.27]. In this case, of the 34 variables selected by RFE for the model, 19 were derived from principal components and corresponding textures, indicating a considerable improvement in using our novel framework to estimate AGB. Together, this study provides a valuable strategy for feature selection to lessen multicollinearity and redundancy.

A Mixed Methods Approach for Fuel Characterisation in Gorse (Ulex europaeus L.) Scrub from High-Density UAV Laser Scanning Point Clouds and Semantic Segmentation of UAV Imagery

The classification and quantification of fuel is traditionally a labour-intensive, costly and often subjective operation, especially in hazardous vegetation types, such as gorse (Ulex europaeus L.) scrub. In this study, unmanned aerial vehicle (UAV) technologies were assessed as an alternative to traditional field methodologies for fuel characterisation. UAV laser scanning (ULS) point clouds were captured, and a variety of spatial and intensity metrics were extracted from these data. These data were used as predictor variables in models describing destructively and non-destructively sampled field measurements of total above ground biomass (TAGB) and above ground available fuel (AGAF). Multiple regression of the structural predictor variables yielded correlations of R2 = 0.89 and 0.87 for destructively sampled measurements of TAGB and AGAF, respectively, with relative root mean square error (RMSE) values of 18.6% and 11.3%, respectively. The best metrics for non-destructive field-measurements yielded correlations of R2 = 0.50 and 0.49, with RMSE values of 40% and 30.8%, for predicting TAGB and AGAF, respectively, indicating that ULS-derived structural metrics offer higher levels of precision. UAV-derived versions of the field metrics (overstory height and cover) predicted TAGB and AGAF with R2 = 0.44 and 0.41, respectively, and RMSE values of 34.5% and 21.7%, demonstrating that even simple metrics from a UAV can still generate moderate correlations. In further analyses, UAV photogrammetric data were captured and automatically processed using deep learning in order to classify vegetation into different fuel categories. The results yielded overall high levels of precision, recall and F1 score (0.83 for each), with minimum and maximum levels per class of F1 = 0.70 and 0.91. In conclusion, these ULS-derived metrics can be used to precisely estimate fuel type components and fuel load at fine spatial resolutions over moderate-sized areas, which will be useful for research, wildfire risk assessment and fuel management operations.

Predicting Subgrade Resistance Value of Hydrated Lime-Activated Rice Husk Ash-Treated Expansive Soil: A Comparison between M5P, Support Vector Machine, and Gaussian Process Regression Algorithms

Resistance value (R-value) is one of the basic subgrade stiffness characterizations that express a material’s resistance to deformation. In this paper, artificial intelligence (AI)-based models—especially M5P, support vector machine (SVM), and Gaussian process regression (GPR) algorithms—are built for R-value evaluation that meets the high precision and rapidity requirements in highway engineering. The dataset of this study comprises seven parameters: hydrated lime-activated rice husk ash, liquid limit, plastic limit, plasticity index, optimum moisture content, and maximum dry density. The available data are divided into three parts: training set (70%), test set (15%), and validation set (15%). The output (i.e., R-value) of the developed models is evaluated using the performance measures coefficient of determination (R2), mean absolute error (MAE), relative squared error (RSE), root mean square error (RMSE), relative root mean square error (RRMSE), performance indicator (ρ), and visual framework (Taylor diagram). GPR is concluded to be the best performing model (R2, MAE, RSE, RMSE, RRMSE, and ρ equal to 0.9996, 0.0258, 0.0032, 0.0012, 0.0012, and 0.0006, respectively, in the validation phase), very closely followed by SVM, and M5P. The application used for the aforementioned approaches for predicting the R-value is also compared with the recently developed artificial neural network model in the literature. The analysis of performance measures for the R-value dataset demonstrates that all the AI-based models achieved comparatively better and reliable results and thus should be encouraged in further research. Sensitivity analysis suggests that all the input parameters have a significant influence on the output, with maximum dry density being the highest.

A potential strategy of carbon dioxide separation using supersonic flows

Carbon capture and storage (CCS) is one of the most promising technologies to tackle climate challenges. The separation of carbon dioxide (CO2) is the key step to achieve the economic and technical objectives of CCS. The present study proposes a potential strategy to separate CO2 using the phase change behavior in supersonic flows, which is not only a clean process of CO2 processing but also provides an efficient way to maximum utilize the thermal energy. To this end, a condensation flow model based on the real gas thermodynamics is developed to obtain an accurate evaluation of the heat and mass transfer due to the homogeneous condensation process of CO2 in supersonic flows. The prediction accuracy of the ideal gas model and real gas model is compared, and the result shows that the real gas condensation model presents a more accurate prediction of CO2 supersonic condensation with the root mean square error (RMSE) of 0.0147. The sensitivity of the two models to inlet pressure is analyzed, which shows that the ideal gas model under-estimated the liquid fractions of CO2 condensation by 2.8% of the total mass as well as over-estimated the latent heat by 20.1% at Wilson point during the heat transfer process. The condensation performances and Wilson point characteristics of CO2 are analyzed by using the real gas model. The prediction model of the relationship between the degree of supercooling, pressure, and expansion rate at the Wilson point was established with the mean relative error of 0.176% and the relative RMSE of 2.275% respectively, which is of great help for further forecasting to obtain the regularity of known data for CO2 separation in supersonic flows.

Comparison of neural network, Gaussian regression, support vector machine, long short-term memory, multi-gene genetic programming, and M5 Trees methods for solving civil engineering problems

In this study, it was investigated that how machine learning (ML) methods show performance in different problems having different characteristics. Six ML approaches including Artificial neural networks (ANN), gaussian process regression (GPR), support vector machine regression (SVMR), long short-term memory (LSTM), multi-gene genetic programming (MGGP) and M5 model tree (M5Tree) were utilized to analyze three independent civil engineering problems belonging to construction management, geotechnical engineering, and hydrological engineering sub-disciplines. Mean absolute percentage error (MAPE), root mean square error (RMSE), coefficient of determination (R2), relative root means square error (RRMSE), Nash–Sutcliffe efficiency (NSE), Kling–Gupta efficiency (KGE), and overall index of model performance (OI) criteria were used to evaluate the performances of the models. Besides performance criteria, the relative performances of the six ML models were assessed using Taylor diagram, Violin diagram and One-Tailed Wilcoxon Signed-Rank Test. For each of the problem considered in this study, the effectiveness of the input parameters on the output parameter has been defined using the Relief Method and Correlation Coefficient. The results show that ANN and MGGP models yielded the most successful estimations for three different problems considered. The best prediction was achieved by MGGP model for hydrological engineering problem. For the construction management, geotechnical engineering problems, the best results were obtained using the ANN model. All models were reliable to solve the geotechnical engineering and hydrological engineering problems while LSTM and SVMR models are not reliable to solve the construction management problem. The most and least effective input parameters on output parameter were contract cost (CC) and work definition number (WDN) for the managerial data set. On the other hand, the most and least effective input parameters on the output parameters for the experimental and natural data sets have been obtained as width of the pile (B), rotation degree (R) and minimum temperature (Tmin), streamflow (Q) data, respectively. The number of data and data selection have a significant effect on the homogeneity of the data set and its representativeness of the problem. The error values obtained in test stage are affected from this condition. The equations to calculate the outputs of each of the problem considered were obtained using MGGP and M5Tree models.

Short-term prediction of particulate matter (PM10 and PM2.5) in Seoul, South Korea using tree-based machine learning algorithms

In this study, highly accurate particulate matter (PM10 and PM2.5) predictions were obtained using meteorological prediction data from the local data assimilation and prediction system (LDAPS) and tree-based machine learning (ML). The study area was Seoul, South Korea, and data from July 2018 to June 2021 as well as LDAPS 36-h predictions with 1-h intervals 4 times a day were used. The predicted PM values were then compared with the observed PM measurements to evaluate the prediction accuracy. The PM prediction performance of the Community Multi-Scale Air Quality (CMAQ)-based chemical transport model (CTM) was compared with that reported by this study. The experimental results report that, among tree-based ML algorithms, light gradient boosting (LGB) is the most suitable for PM prediction. The PM prediction results of the LGB algorithm for the hourly test data were: bias = −0.10 μg/m3, root mean square error (RMSE) = 13.15 μg/m3, and R2 = 0.86 for PM10 and bias = −0.02 μg/m3, RMSE = 7.48 μg/m3, and R2 = 0.83 for PM2.5, and for daily mean were: RMSE ≤1.16 μg/m3 and R2 = 0.996. The relative RMSE (%RMSE) is 21% lower than the results of the CTM model, and R2 is 0.20 higher. Even in the high PM concentration case prediction results, the algorithm showed good predictive performance with %RMSE = 8.91%–20.43% and R2 = 0.89–0.97. Therefore, in addition to the CTM, high-accuracy PM prediction results using ML can also be used for air quality monitoring and improvement.

Examining the Role of UAV Lidar Data in Improving Tree Volume Calculation Accuracy

Traditional forest inventories are based on field surveys of established sample plots, which involve field measurements of individual trees within a sample plot and the selection of proper allometric equations for tree volume calculation. Thus, accurate field measurements and properly selected allometric equations are two crucial factors for providing high-quality tree volumes. One key problem is the difficulty in accurately acquiring tree height data, resulting in high uncertainty in tree volume calculation when the diameter at breast height (DBH) alone is used. This study examined the uncertainty of tree height measurements using different means and the impact of allometric models on tree volume estimation accuracy. Masson pine and eucalyptus plantations in Fujian Province, China, were selected as examples; their tree heights were measured three ways: using an 18-m telescopic pole, UAV Lidar (unmanned aerial vehicle, light detection and ranging) data, and direct measurement of felled trees, with the latest one as a reference. The DBH-based and DBH–height-based allometric equations corresponding to specific tree species were used for the calculations of tree volumes. The results show that (1) tree volumes calculated from the DBH-based models were lower than those from the DBH–height-based models. On average, tree volumes were underestimated by 0.018 m3 and 0.117 m3 for Masson pine and eucalyptus, respectively, while the relative root-mean-squared errors (RMSEr) were 24.04% and 33.90%, respectively, when using the DBH-based model; (2) the tree height extracted from UAV Lidar data was more accurate than that measured using a telescopic pole, because the pole measurement method generally underestimated the tree height, especially when the trees were taller than the length of the pole (18 m in our study); (3) the tree heights measured using different methods greatly impacted the accuracies of tree volumes calculated using the DBH–height model. The telescopic-pole-measured tree heights resulted in a relative error of 9.1–11.8% in tree volume calculations. This research implies that incorporation of UAV Lidar data with DBH field measurements can effectively improve tree volume estimation and could be a new direction for sample plot data collection in the future.

Online 2D Fluorescence Monitoring in Microtiter Plates Allows Prediction of Cultivation Parameters and Considerable Reduction in Sampling Efforts for Parallel Cultivations of Hansenula polymorpha

Multi-wavelength (2D) fluorescence spectroscopy represents an important step towards exploiting the monitoring potential of microtiter plates (MTPs) during early-stage bioprocess development. In combination with multivariate data analysis (MVDA), important process information can be obtained, while repetitive, cost-intensive sample analytics can be reduced. This study provides a comprehensive experimental dataset of online and offline measurements for batch cultures of Hansenula polymorpha. In the first step, principal component analysis (PCA) was used to assess spectral data quality. Secondly, partial least-squares (PLS) regression models were generated, based on spectral data of two cultivation conditions and offline samples for glycerol, cell dry weight, and pH value. Thereby, the time-wise resolution increased 12-fold compared to the offline sampling interval of 6 h. The PLS models were validated using offline samples of a shorter sampling interval. Very good model transferability was shown during the PLS model application to the spectral data of cultures with six varying initial cultivation conditions. For all the predicted variables, a relative root-mean-square error (RMSE) below 6% was obtained. Based on the findings, the initial experimental strategy was re-evaluated and a more practical approach with minimised sampling effort and elevated experimental throughput was proposed. In conclusion, the study underlines the high potential of multi-wavelength (2D) fluorescence spectroscopy and provides an evaluation workflow for PLS modelling in microtiter plates.

Artificial intelligence-based software (AID-FOREST) for tree detection: A new framework for fast and accurate forest inventorying using LiDAR point clouds

Forest inventories are essential to accurately estimate different dendrometric and forest stand parameters. However, classical forest inventories are time consuming, slow to conduct, sometimes inaccurate and costly. To address this problem, an efficient alternative approach has been sought and designed that will make this type of field work cheaper, faster, more accurate, and easier to complete. The implementation of this concept has required the development of a specifically designed software called “Artificial Intelligence for Digital Forest (AID-FOREST)”, which is able to process point clouds obtained via mobile terrestrial laser scanning (MTLS) and then, to provide an array of multiple useful and accurate dendrometric and forest stand parameters. Singular characteristics of this approach are: No data pre-processing is required either pre-treatment of forest stand; fully automatic process once launched; no limitations by the size of the point cloud file and fast computations. To validate AID-FOREST, results provided by this software were compared against the obtained from in-situ classical forest inventories. To guaranty the soundness and generality of the comparison, different tree species, plot sizes, and tree densities were measured and analysed. A total of 76 plots (10,887 trees) were selected to conduct both a classic forest inventory reference method and a MTLS (ZEB-HORIZON, Geoslam, ltd.) scanning to obtain point clouds for AID-FOREST processing, known as the MTLS-AIDFOREST method. Thus, we compared the data collected by both methods estimating the average number of trees and diameter at breast height (DBH) for each plot. Moreover, 71 additional individual trees were scanned with MTLS and processed by AID-FOREST and were then felled and divided into logs measuring 1 m in length. This allowed us to accurately measure the DBH, total height, and total volume of the stems. When we compared the results obtained with each methodology, the mean detectability was 97% and ranged from 81.3 to 100%, with a bias (underestimation by MTLS-AIDFOREST method) in the number of trees per plot of 2.8% and a relative root-mean-square error (RMSE) of 9.2%. Species, plot size, and tree density did not significantly affect detectability. However, this parameter was significantly affected by the ecosystem visual complexity index (EVCI). The average DBH per plot was underestimated (but was not significantly different from 0) by the MTLS-AIDFOREST, with the average bias for pooled data being 1.8% with a RMSE of 7.5%. Similarly, there was no statistically significant differences between the two distribution functions of the DBH at the 95.0% confidence level. Regarding the individual tree parameters, MTLS-AIDFOREST underestimated DBH by 0.16 % (RMSE = 5.2 %) and overestimated the stem volume (Vt) by 1.37 % (RMSE = 14.3 %, although the BIAS was not statistically significantly different from 0). However, the MTLS-AIDFOREST method overestimated the total height (Ht) of the trees by a mean 1.33 m (5.1 %; relative RMSE = 11.5 %), because of the different height concepts measured by both methodological approaches. Finally, AID-FOREST required 30 to 66 min per ha−1 to fully automatically process the point cloud data from the *.las file corresponding to a given hectare plot. Thus, applying our MTLS-AIDFOREST methodology to make full forest inventories, required a 57.3 % of the time required to perform classical plot forest inventories (excluding the data postprocessing time in the latter case). A free trial of AID-FOREST can be requested at [email protected]

A case study of NeuralProphet and nonlinear evaluation for high accuracy prediction in short-term forecasting in PV solar plant

Prediction of the energy, active production from the PV solar plant is a challenge in cloudy weather or with clouds over the solar plant; therefore, it has impact in the planning of the power system, especially in the season analysis and prediction accuracy adjustments, for example in holidays. In 2022, some authors published some analysis associated to horizontal pyranometers and the limits in the evaluation of the data, the Mean Bias Error of Daily Solar Irradiation average (MIEave) ranges from 0.17% to 2.86% associated a sudden change in the weather, it increases the “risk of misestimating the potential electricity generation” with short-term error of more than 50% and the Global Horizontal Irradiance (GHI) has a mean bias error (MBE) of at least ±8% [1]. In this research article, a novel proposal for short-term forecasting combines the satellite with meteorological station data and statistical model associated to the new seasonality analysis by using two approaches: i) NeuralProphet, Ridge regression, ii) Long Short-Term Memory with convolutional neural networks. Besides, it requires three KPI as feedback, it is the mean absolute error (MAE), relative Root mean square error (RMSE), and mean absolute percentage error (MAPE). The results demonstrate a MAPE of 5.93% and a computational time 852.10 s and the comparison with new predictions methods from 2019 to 2021. This research article illustrates the new approach with the forecasting method in a case of the PV solar plant in Peru and proves the robustness and seasonality results, and new short-terms improvements associated to external influence as cloudy conditions and resource availability. Our findings are an improvement of the model MAPE 12.14%–5.93%; even compared with the literature and currently models as ARIMA-LSTM with 10.57%, LSTM with NN and G, SARIMA and SVM considering Gaussian White Noise with 8.14% and Prophet with SVM with 8.81%.

Benchmarking laser scanning and terrestrial photogrammetry to extract forest inventory parameters in a complex temperate forest

National forest inventories (NFI) are important for the assessment of the state and development of forests. Traditional NFIs often rely on statistical sampling approaches as well as expert assessment which may suffer from observer bias and may lack robustness for time series analysis. Over the course of the last decade, close-range remote sensing techniques such as terrestrial and mobile laser scanning became ever more established for the assessment of three-dimensional (3D) forest structure. With the ongoing trend to make the systems smaller, easier to use and more efficient, the pathway is being opened for an operational inclusion of such devices within the framework of an NFI to support the traditional field assessment. Close-range remote sensing could potentially speed up field inventory work as well as increase the area in which certain parameters are assessed. Benchmarks are needed to evaluate the performance of different close-range remote sensing devices and approaches, both in terms of efficiency as well as accuracy. In this study we evaluate the performance of two terrestrial (TLS), one handheld mobile (PLS) and two drone based (UAVLS) laser scanning systems to detect trees and extract the diameter at breast height (DBH) in three plots with a steep gradient in tree and understorey vegetation density. As a novelty, we also tested the acquisition of 3D point-clouds using a low-cost action camera (GoPro) in conjunction with the Structure from Motion (SfM) technique and compared its performance with those of the more costly LiDAR devices. Among the many parameters evaluated in traditional NFIs, the focus of the performance evaluation of this study is set on the automatic tree detection and DBH extraction. The results showed that TLS delivers the highest tree detection rate (TDR) of up to 94.6% under leaf-off and up to 82% under leaf-on conditions and a relative RMSE (rRMSE) for the DBH extraction between 2.5 and 9%, depending on the undergrowth complexity. The tested PLS system (leaf-on) achieved a TDR of up to 80% with an rRMSE between 3.7 and 5.8%. The tested UAVLS systems showed lowest TDR of less than 77% under leaf-off and less than 37% under leaf-on conditions. The novel GoPro approach achieved a TDR of up to 53% under leaf-on conditions. The reduced TDR can be explained by the reduced area coverage due to the chosen circular acquisition path taken with the GoPro approach. The DBH extraction performance on the other hand is comparable to those of the LiDAR devices with an rRMSE between 2 and 9%. Further benchmarks are needed in order to fully assess the applicability of these systems in the framework of an NFI. Especially the robustness under varying forest conditions (seasonality) and over a broader range of forest types and canopy structure has to be evaluated.

Adaptiveness of RGB-image derived algorithms in the measurement of fractional vegetation coverage

BackgroundFractional vegetation coverage (FVC) is a crucial parameter in determining vegetation structure. Automatic measurement of FVC using digital images captured by mobile smart devices is a potential direction for future research on field survey methods in plant ecology, and this algorithm is crucial for accurate FVC measurement. However, there is a lack of insight into the influence of illumination on the accuracy of FVC measurements. Therefore, the main objective of this research is to assess the adaptiveness and performance of different algorithms under varying light conditions for FVC measurements and to deepen our understanding of the influence of illumination on FVC measurement.Methods and resultsBased on a literature survey, we selected four algorithms that have been reported to have high accuracy in automatic FVC measurements. The first algorithm (Fun01) identifies green plants based on the combination of R/G, B/G, and ExG (R, G, and B are the actual pixel digital numbers from the images based on each RGB channel, ExG is the abbreviation of the Excess Green index), the second algorithm (Fun02) is a decision tree that uses color properties to discriminate plants from the background, the third algorithm (Fun03) uses ExG-ExR (ExR is the abbreviation of the Excess Red index) to recognize plants in the image, and the fourth algorithm (Fun04) uses ExG and O{text{tsu}} to separate the plants from the background. Otsu is an algorithm used to determine a threshold to transform the image into a binary image for the vegetation and background. We measured the FVC of several surveyed quadrats using these four algorithms under three scenarios, namely overcast sky, solar forenoon, and solar noon. FVC values obtained using the Photoshop-assisted manual identification method were used as a reference to assess the accuracy of the four algorithms selected. Results indicate that under the overcast sky scenario, Fun01 was more accurate than the other algorithms and the MAPE (mean absolute percentage error), BIAS, relBIAS (relative BIAS), RMSE (root mean square error), and relRMSE (relative RMSE) are 8.68%, 1.3, 3.97, 3.13, and 12.33%, respectively. Under the scenario of the solar forenoon, Fun02 (decision tree) was more accurate than other algorithms, and the MAPE, BIAS, relBIAS, RMSE, and relRMSE are 22.70%, − 2.86, − 7.70, 5.00, and 41.23%. Under the solar noon scenario, Fun02 was also more accurate than the other algorithms, and the MAPE, BIAS, relBIAS, RMSE, and relRMSE are 20.60%, − 6.39, − 20.67, 7.30, and 24.49%, respectively.ConclusionsGiven that each algorithm has its own optimal application scenario, among the four algorithms selected, Fun01 (the combination of R/G, B/G, and ExG) can be recommended for measuring FVC on cloudy days. Fun02 (decision tree) is more suitable for measuring the FVC on sunny days. However, it considerably underestimates the FVC in most cases. We expect the findings of this study to serve as a useful reference for automatic vegetation cover measurements.

Optimizing Deficit Irrigation Management to Improve Water Productivity of Greenhouse Tomato under Plastic Film Mulching Using the RZ-SHAW Model

Deficit irrigation (DI) is a widely recognized water-saving irrigation method, but it is difficult to precisely quantify optimum DI levels in tomato production. In this study, the Root Zone Water Quality-Simultaneous Heat and Water (RZ-SHAW) model was used to evaluate the potential effects of different DI levels on tomato growth in a drip-irrigated field. Combinations of five DI scenarios were tested in greenhouse field experiments under plastic film mulching according to the percentage of crop evapotranspiration (ET), i.e., ET50, ET75, ET100, ET125, and ET150. The model was calibrated by using the ET100 scenario, and validated with four other scenarios. The simulation results showed that the predictions of tomato growth parameters and soil water were in good agreement with the observed data. The relative root mean square error (RRMSE), the percent bias (PBIAS), index of agreement (IoA) and coefficient of determination (R2) for leaf area index (LAI), plant height and soil volumetric water content (VWC) along the soil layers were &lt;23.5%, within ±16.7%, &gt;0.72 and &gt;0.56, respectively. The relative errors (REs) of simulated biomass and yield were 3.5–8.7% and 7.0–14.0%, respectively. There was a positive correlation between plant water stress factor (PWSF) and DI levels (p &lt; 0.01). The calibrated model was subsequently run with 45 different DI scenarios from ET0 to ET225 to explore optimal DI management for maximizing water productivity (WP) and yield. It was found that the maximum WP and yield occurred in ET95 and ET200, with values of 28.3 kg/(ha·mm) and 7304 kg/ha, respectively. The RZ-SHAW demonstrated its capacity to evaluate the effects of DI management on tomato growth under plastic film mulching. The parameterized model can be used to optimize DI management for improving WP and yield based on the water stress-based method.