Accurate Volume Estimation Research Articles

Human-Mimetic Estimation of Food Volume from a Single-View RGB Image Using an AI System.

It is well known that many chronic diseases are associated with unhealthy diet. Although improving diet is critical, adopting a healthy diet is difficult despite its benefits being well understood. Technology is needed to allow an assessment of dietary intake accurately and easily in real-world settings so that effective intervention to manage being overweight, obesity, and related chronic diseases can be developed. In recent years, new wearable imaging and computational technologies have emerged. These technologies are capable of performing objective and passive dietary assessments with a much simplified procedure than traditional questionnaires. However, a critical task is required to estimate the portion size (in this case, the food volume) from a digital image. Currently, this task is very challenging because the volumetric information in the two-dimensional images is incomplete, and the estimation involves a great deal of imagination, beyond the capacity of the traditional image processing algorithms. In this work, we present a novel Artificial Intelligent (AI) system to mimic the thinking of dietitians who use a set of common objects as gauges (e.g., a teaspoon, a golf ball, a cup, and so on) to estimate the portion size. Specifically, our human-mimetic system “mentally” gauges the volume of food using a set of internal reference volumes that have been learned previously. At the output, our system produces a vector of probabilities of the food with respect to the internal reference volumes. The estimation is then completed by an “intelligent guess”, implemented by an inner product between the probability vector and the reference volume vector. Our experiments using both virtual and real food datasets have shown accurate volume estimation results.

Prediction of Liver Volume from Liver Transplant Donor using Biometric Formula compared with Computed Topography Volumetry

Introduction. Liver volume calculation is critical in assessing the compatibility and resectability of the graft in living donor liver transplants (LDLT). An accurate estimation of liver volume is a predictor for successful LDLT. The gold standard of liver volume estimation is CT Volumetry. Despite several limitations in the availability of software, facility, and time consumed, there is still disagreement of biometric formula to predict liver volume in Indonesia.

Intracerebral hemorrhage volume estimation: Is modification of the ABC/2 formula necessary according to the hematoma shape?

ObjectiveWe studied the proposal to modify the ABC/2 formula to ABC/3 for irregular-shaped intracerebral hematoma (ICH) volume estimation. Patients and methodsThe volume of 133 ICHs were estimated with Kwak’s (simplified C; all slices with hemorrhage are considered equal), Kothari’s (weighted C) and coronal (reformatted C; measuring C directly on coronal reformatted images) ABC/2 methods, and compared with computer-assisted planimetric measurements. The accuracy, precision and correlation of three ABC/2 methods and their ABC/3 modifications were determined in smooth (Barras’ group 1 or 2) and irregular (Barras’ group 3–5) shaped ICHs. ResultsAs the hematoma size increases, the shape becomes irregular. In all hematomas, both smooth (n = 81) and irregular (n = 52) shaped, Kothari's ABC/2 formula provided the closest result to the planimetric measurement, with an underestimation of 1.77 mL, and 10.2% difference on average. Kothari’s ABC/2 disclosed the best correlation (Lin’s coefficient=0.9622) regardless of ICH shape. When simplified-ABC/2 method was modified as ABC/3, volume estimation accuracy increased (Correlation coefficient increased from 0.838 to 0.915) for irregular hematomas; however, despite this improvement the accuracy remained below the Kothari’s ABC/2 (not ABC/3) method. Neither reformatted coronal ABC/2 nor its ABC/3 modification provided any advantage over ABC/x formulas with slice counting. ConclusionKothari's ABC/2 method is a valid method for estimation of ICH volume for both regular and irregular shaped hematomas. Simplified (Kwak's) ABC/2 or coronal ABC/2, or their /3 counterparts do not provide additional advantage.

Preoperative volume estimation in transverse upper gracilis flap surgery: A pilot study

The transverse upper gracilis (TUG) flap provides a good alternative to the gold standard DIEP in breast reconstruction. However, flap volume estimates are subjective, making preoperative planning potentially challenging. To derive a reliable, accurate, and reproducible mathematical algorithm for the preoperative calculation of TUG flap volumes. Nineteen consecutive patients with 30 TUG flaps were prospectively included. On the assumption that the TUG flap resembles two isosceles prisms, the formula of the volume of a prism was used to calculate their preoperative flap weights. These were then intraoperatively compared to the actual flap weights. A regression equation was calculated from the correlation analysis of 10 random flaps. This was then applied to the remaining 20 flaps to assess for improved reliability and weight prediction accuracy. The prism volume equation used to clinically calculate flap volumes was: Geometric flap weight = (h1bT)/2+ (h2bT)/2, (h = height, b = base, T = flap thickness); all in centimetres. Geometric and actual flap weights were found to be significantly correlated (r2 = 0.977) generating the following regression formula: predicted TUG weight = 0.924 × geometric weight + 26.601. When this was applied to the remaining 20 flaps, no significant difference was found (p = 0.625) between predicted and actual flap weights, demonstrating an increased accuracy of predicting flap volume. The proposed formula provides the clinician with a more accurate and reliable estimation of available TUG flap volume and may potentially aid with preoperative planning and patient consultations.

A neural network–based approach for fill factor estimation and bucket detection on construction vehicles

AbstractBucket fill factor is of paramount importance in measuring the productivity of construction vehicles, which is the percentage of materials loaded in the bucket within one scooping. Additionally, the locational information of the bucket is also indispensable for scooping trajectory planning. Some research has been conducted to measure it via state‐of‐the‐art computer vision approaches, but their robustness against various environment conditions is not considered. The aim of this study is to fill this gap and six distinctive environment settings are included. Images captured by a stereo camera are used to generate point clouds before being structured into 3D maps. This novel preprocessing pipeline for deep learning is originally proposed and its feasibility has been validated through this study. Moreover, multitask learning is employed to exploit the positive relationship among two tasks: fill factor prediction and bucket detection. Therefore, after preprocessing, 3D maps are forwarded to a faster region with convolutional neural network incorporated with an improved residual neural network. The value of fill factor is acquired via a classification and probabilistic–based approach, which is novel, achieving an inspiring result (overall volume estimation accuracy: 95.23% and detection precision: 92.62%) at the same time.

Development of a simplified method for optimally sizing hot water storage tanks subject to short-term intermittent charge/discharge cycles

Hot water storage tanks play a pivotal role in the transition to low carbon energy systems by increasing energy efficiency and allowing a range of non-conventional intermittent energy sources in the energy mix such as renewable energy and recovered waste heat. Apart from lowering greenhouse gas emissions, sizing these systems appropriately can lead to significant reductions in annual energy cost. In the current study, a simple analytical method is developed which can be employed by a user to optimally size hot water storage tanks using only the system residual heating profile and a limited number of system characteristics. The proposed method is based on the results of an analysis combining numerical simulation and mathematical programming techniques. A transient numerical model (TRNSYS) of a hot water storage tank system is developed to simulate randomly generated residual heating profiles under 36 distinct scenarios. For each residual heating profile, the optimal storage tank volume is determined by coupling the TRNSYS model to an optimization framework developed in Matlab®. Fast Fourier transform techniques are used to decompose each residual heating profile, and ordinary least-squares regression models are used to relate the residual heating profile's dominant amplitudes and periods to the optimal storage volume. Users of the proposed method can quickly obtain an accurate estimate of the optimal hot water storage tank volume for a given system simply by inputting the following data into a custom-developed analytical expression and solving directly: the dominant amplitudes and periods of the residual heating profile (determined by the user), and the regression model parameters corresponding to the scenario that most closely matches the given system's characteristics (obtained in the current study). Thus, no specialized modelling software is required on the user's behalf. Results show that all scenarios with a source temperature of 95°C, a source to load temperature difference of at least 35°C, and an auxiliary energy price of at least 0.105 USD/kWh have associated R2 values of 0.8 or greater, indicating the proposed sizing method is able to predict the optimal storage tank volume with a high degree of confidence. A case study is presented to demonstrate the proposed sizing method for a hypothetical solar-thermal domestic hot water system in Montréal, Canada.

Occluded Grape Cluster Detection and Vine Canopy Visualisation Using an Ultrasonic Phased Array.

Grape yield estimation has traditionally been performed using manual techniques. However, these tend to be labour intensive and can be inaccurate. Computer vision techniques have therefore been developed for automated grape yield estimation. However, errors occur when grapes are occluded by leaves, other bunches, etc. Synthetic aperture radar has been investigated to allow imaging through leaves to detect occluded grapes. However, such equipment can be expensive. This paper investigates the potential for using ultrasound to image through leaves and identify occluded grapes. A highly directional low frequency ultrasonic array composed of ultrasonic air-coupled transducers and microphones is used to image grapes through leaves. A fan is used to help differentiate between ultrasonic reflections from grapes and leaves. Improved resolution and detail are achieved with chirp excitation waveforms and near-field focusing of the array. The overestimation in grape volume estimation using ultrasound reduced from 222% to 112% compared to the 3D scan obtained using photogrammetry or from 56% to 2.5% compared to a convex hull of this 3D scan. This also has the added benefit of producing more accurate canopy volume estimations which are important for common precision viticulture management processes such as variable rate applications.

Validation and estimation of spleen volume via computer-assisted segmentation on clinically acquired CT scans.

Purpose: Deep learning is a promising technique for spleen segmentation. Our study aims to validate the reproducibility of deep learning-based spleen volume estimation by performing spleen segmentation on clinically acquired computed tomography (CT) scans from patients with myeloproliferative neoplasms. Approach: As approved by the institutional review board, we obtained 138 de-identified abdominal CT scans. A sum of voxel volume on an expert annotator's segmentations establishes the ground truth (estimation 1). We used our deep convolutional neural network (estimation 2) alongside traditional linear estimations (estimation 3 and 4) to estimate spleen volumes independently. Dice coefficient, Hausdorff distance, coefficient, Pearson coefficient, the absolute difference in volume, and the relative difference in volume were calculated for 2 to 4 against the ground truth to compare and assess methods' performances. We re-labeled on scan-rescan on a subset of 40 studies to evaluate method reproducibility. Results: Calculated against the ground truth, the coefficients for our method (estimation 2) and linear method (estimation 3 and 4) are 0.998, 0.954, and 0.973, respectively. The Pearson coefficients for the estimations against the ground truth are 0.999, 0.963, and 0.978, respectively (paired -tests produced between 2 and 3, and 2 and 4). Conclusion: The deep convolutional neural network algorithm shows excellent potential in rendering more precise spleen volume estimations. Our computer-aided segmentation exhibits reasonable improvements in splenic volume estimation accuracy.

Accurate Estimation of Total Intracranial Volume in MRI using a Multi-tasked Image-to-Image Translation Network.

Total intracranial volume (TIV) is the volume enclosed inside the cranium, inclusive of the meninges and the brain. TIV is extensively used to correct variations in inter-subject head size for the evaluation of neurodegenerative diseases. In this work, we present an automatic method to generate a TIV mask from MR images while synthesizing a CT image to be used in subsequent analysis. In addition, we propose an alternative way to obtain ground truth TIV masks using a semi-manual approach, which results in significant time savings. We train a conditional generative adversarial network (cGAN) using 2D MR slices to realize our tasks. The quantitative evaluation showed that the model was able to synthesize CT and generate TIV masks that closely approximate the reference images. This study also provides a comparison of the described method against skull stripping tools that output a mask enclosing the cranial volume, using MRI scan. In particular, highlighting the deficiencies in using such tools to approximate the volume using MRI scan.

Analysis of Taper Functions for Larix olgensis Using Mixed Models and TLS

Terrestrial laser scanning (TLS) plays a significant role in forest resource investigation, forest parameter inversion and tree 3D model reconstruction. TLS can accurately, quickly and nondestructively obtain 3D structural information of standing trees. TLS data, rather than felled wood data, were used to construct a mixed model of the taper function based on the tree effect, and the TLS data extraction and model prediction effects were evaluated to derive the stem diameter and volume. TLS was applied to a total of 580 trees in the nine larch (Larix olgensis) forest plots, and another 30 were applied to a stem analysis in Mengjiagang. First, the diameter accuracies at different heights of the stem analysis were analyzed from the TLS data. Then, the stem analysis data and TLS data were used to establish the stem taper function and select the optimal basic model to determine a mixed model based on the tree effect. Six basic models were fitted, and the taper equation was comprehensively evaluated by various statistical metrics. Finally, the optimal mixed model of the plot was used to derive stem diameters and trunk volumes. The stem diameter accuracy obtained by TLS was >98%. The taper function fitting results of these data were approximately the same, and the optimal basic model was Kozak (2002)-II. For the tree effect, a6 and a9 were used as the mixed parameters, the mixed model showed the best fit, and the accuracy of the optimal mixed model reached 99.72%.The mixed model accuracy for predicting the tree diameter was between 74.22% and 97.68%, with a volume estimation accuracy of 96.38%. Relative height 70 (RH70) was the optimum height for extraction, and the fitting accuracy of the mixed model was higher than that of the basic model.

Using matrix-induced ion suppression combined with LC-MS/MS for quantification of trimethylamine-N-oxide, choline, carnitine and acetylcarnitine in dried blood spot samples

Recently, there has been significant interest in the influences of the human gut microbiota on many diseases, such as cardiovascular disease (CVD) and metabolic disorders. Trimethylamine N-oxide (TMAO) is one of the most frequently discussed gut-derived metabolites. Dried blood spot (DBS) sampling has been regarded as an attractive alternative sampling strategy for clinical studies and offers many advantages. For DBS sample processing, whole-spot analysis could minimize hematocrit-related bias, but it requires blood volume calibration. This study developed a method combining matrix-induced ion suppression (MIIS) with liquid chromatography-electrospray ionization-tandem mass spectrometry (LC-ESI-MS/MS) to estimate blood volume and quantify TMAO and its precursors and derivatives, including choline, carnitine and acetylcarnitine, in DBSs. The MIIS method used an ion suppression indicator (ISI) to measure the extent of ion suppression caused by the blood matrix, which was related to the blood volume. The results showed that the volume estimation accuracy of the MIIS method was within 91.7–109.7%. The combined MIIS and LC-MS/MS method for quantifying TMAO, choline, carnitine and acetylcarnitine was validated in terms of linearity, precision and accuracy. The quantification accuracy was within 91.2–113.2% (with LLOQ <119%), and the imprecision was below 8.0% for all analytes. A stability study showed that the analytes in DBSs were stable at all evaluated temperatures for at least 30 days. The validated method was applied to quantify DBS samples (n = 56). Successful application of the new method demonstrated the potential of this method for real-world DBS samples and to facilitate our understanding of the gut microbiota in human health.

Functional Segmentation for Preoperative Liver Resection Based on Hepatic Vascular Networks

Successful liver resection relies on accurate estimation of future liver remnant volume (FLRV). According to Couinaud’s scheme, a liver is composed of eight functionally independent segments, each of which has its own vascular in- and out-flow tracks. Segmenting a liver by this scheme is vital to postoperative regeneration and hence prognosis outcome. Conventionally, estimation of liver segments was often done by hand on 3D computed tomography. The process is generally tedious, time consuming, and prone to observer variability. Alternatively, computerized methods had been proposed but impeded by anatomically irrelevant approximation and manually specified markers. To resolve the issues, this paper presents a novel method for functional liver segmentation. Its main contribution was performing analyses of differential geometry directly on a liver surface and interior venous system. Except for a few points being placed on major vessels, anatomical references required for defining all separating surfaces were automatically identified. To demonstrate its merits, virtual liver resection was implemented on the standard MICCAI SLIVER07 dataset, and the resultant segments were benchmarked against four most related works. Visual and numerical assessments reported herein indicated that our method could faithfully label all Couinaud’s segments, especially the caudate, with lesser degree of user interaction. The preliminary findings suggested that it can be integrated into augmented surgical planning and intervention.

Three-dimensional (3D) Models Based on Pre-operative Computed Tomography Scans to Personalize the Planned Liver Surgery in Complex Cases

Purpose: In the last two decades, an increased number of patients has been submitted to liver resection for malignancy. That was possible thank to a better knowledge of tumor biology, perioperative management as well as improvement of surgical technique. Recently, three-dimensional (3D) models visualization systems based on pre-operative imaging has been proposed to plan liver surgery in complex cases. The aim of this study was to assess the feasibility and accuracy of a 3D models for patients undergoing major/extended liver resection for malignancies. Method: The Medical Device Custom Made with Hyper Accuracy 3DTM reconstruction method was applied to preoperative computed tomography (CT) scans of patients undergoing major/extended liver resections for malignancies at our institution since February 2020. The 3D models were used to guide the liver resections, including details regarding anatomical variations of the main vessels (hepatic artery, hepatic veins and portal vein) and the bile ducts as well as their relationship with the tumor and an accurate estimation of the future remnant liver volume (FRLV). Results: A total of 17 patients were included In this preliminary study. The most frequent diagnoses were hilar cholangiocarcinoma (PH-CCA, 52.9%) and intrahepatic cholangiocarcinoma (I-CCA, 23.5%). The 3D reconstruction showed anatomical vessels’ variations in 12 (70.6%) patients that were confirmed intraoperatively. A total of 5 (29.4%) patients underwent portal resection for tumor invasion. A good to optima correspondence between the hepatectomy planned using the 3D models and the intraoperative data was found in terms of bile duct and vessels resected. With a median FRLV at the 3D reconstruction of 31%, 4 (23.5%) patients had transient post hepatectomy liver failure (PHLF) of grade A/B. The median estimated blood loss was 750 mL and the median operative time was 532 minute while the median length of stay was 12 days. In 2 (11.7%) patients the resection planned on the basis of the CT imaging was changed after the pre-operative analysis of 3D reconstructions. Conclusion: These preliminary data regarding the use of 3D models in the accurate and personalize planning of complex liver resections showed that the 3D liver reconstruction might provide useful and additional information about anatomical vessels’ variations and tumor extension. The analysis of 3D reconstructions might lead the surgeon to change the planned resection.

Deep Learning-Based Auto-Segmentation and Evaluation of Vallecular Residue in Videofluoroscopy

Background: Pharyngeal residue is widely considered as an indicator of swallow impairment in videofluoroscopic studies. The volume of residue indicates potential risk for penetration-aspiration on subsequent swallows while the accuracy and measurement guidelines of residue volume estimation varies significantly among human judges and facilities. We aimed to use deep learning to reliably and efficiently estimate the residue volume in vallecular and classify the potential risk alert of penetration to clinicians. Methods: Patients with suspected dysphagia were recruited between 2015 to 2018. We selected 172 qualified residue samples from 3142 swallows. We first developed deep learning algorithm that can efficiently identify and segment the residue area remaining in vallecula. We split our dataset into three parts, with train, test1 (validation) and test2. In the training phase, we deployed transfer learning techniques and ensemble strategies on train dataset. The model was finally determined by its performance on validation dataset. The segmentation performance of this architecture was evaluated on test 2. We also derived $\%(C2-C4)^2$ residue scale using results from our segmentation algorithm and physiological measurement from our previous study. Findings: We tested the reliability among human raters. Testing on a held-out set of swallow cases, ensemble architecture achieved a dice coefficient of 0.72 and 0.70 of segment accuracy on segmentation, and a sensitivity of 0.92 in reference to $\%(C2-C4)^2$ measure scale. Interpretation: These findings demonstrate that automatic deep learning methods hold tremendous promise for improving objectivity in post-swallow residue evaluation, which has historically be a challenge due to carriable training and measurement methods. Funding: NIH Declaration of Interests: The authors declare no competing interests. Ethics Approval Statement: The Institutional Review Board at the University of Pittsburgh approved the protocol of this study and all participants were informed and signed the consent.

Estimating Surgical Blood Loss Volume Using Continuously Monitored Vital Signs.

Background: There are currently no effective and accurate blood loss volume (BLV) estimation methods that can be implemented in operating rooms. To improve the accuracy and reliability of BLV estimation and facilitate clinical implementation, we propose a novel estimation method using continuously monitored photoplethysmography (PPG) and invasive arterial blood pressure (ABP). Methods: Forty anesthetized York Pigs (31.82 ± 3.52 kg) underwent a controlled hemorrhage at 20 mL/min until shock development was included. Machine-learning-based BLV estimation models were proposed and tested on normalized features derived by vital signs. Results: The results showed that the mean ± standard deviation (SD) for estimating BLV against the reference BLV of our proposed random-forest-derived BLV estimation models using PPG and ABP features, as well as the combination of ABP and PPG features, were 11.9 ± 156.2, 6.5 ± 161.5, and 7.0 ± 139.4 mL, respectively. Compared with traditional hematocrit computation formulas (estimation error: 102.1 ± 313.5 mL), our proposed models outperformed by nearly 200 mL in SD. Conclusion: This is the first attempt at predicting quantitative BLV from noninvasive measurements. Normalized PPG features are superior to ABP in accurately estimating early-stage BLV, and normalized invasive ABP features could enhance model performance in the event of a massive BLV.

Improving wood volume predictions in dry tropical forest in the semi-arid Brazil

The volumetric variability of dry tropical forests in Brazil and the scarcity of studies on the subject show the need for the development of techniques that make it possible to obtain adequate and accurate wood volume estimates. In this study, we analyzed a database of thinning trees from a forest management plan in the Contendas de Sincorá National Forest, southwestern Bahia State, Brazil. The data set included a total of 300 trees with a trunk diameter ranging from 5 to 52 cm. Adjustments, validation and statistical selection of four volumetric models were performed. Due to the difference in height values for the same diameter and the low correlation between both variables, we do not suggest models which only use the diameter at breast height (DBH) variable as a predictor because they accommodate the largest estimation errors. In comparing the best single entry model (Hohenald-Krenn) with the Spurr model (best fit model), it is noted that the exclusion of height as a predictor causes the values of 136.44 and 0.93 for Akaike information criterion (AIC) and adjusted determination coefficient (R 2adj ), which are poorer than the second best model (Schumacher-Hall). Regarding the minimum sample size, errors in estimation (root mean square error (RMSE) and bias) of the best model decrease as the sample size increases, especially when a larger number of trees with DBH ≥ 5.0 cm are randomly sampled. Stratified sampling by diameter class produces smaller volume prediction errors than random sampling, especially when considering all trees. In summary, the Spurr and Schumacher-Hall models perform better. These models suggest that the total variance explained in the estimates is not less than 95%, producing reliable forecasts of the total volume with shell. Our estimates indicate that the bias around the average is not greater than 7%. Our results support the decision to use regression methods to build models and estimate their parameters, seeking stratification strategies in diameter classes for the sample trees. Volume estimates with valid confidence intervals can be obtained using the Spurr model for the studied dry forest. Stratified sampling of the data set for model adjustment and selection is necessary, since we find significant results with mean error square root values and bias of up to 70% of the total database.

Is it possible to estimate volume of bone defects formed on dry sheep mandibles more practically by secondarily reconstructing section thickness of cone beam computed tomography images?

The purpose of this study was to evaluate the effect of section thickness on volume estimations of bone defects scanned using cone beam computed tomography (CBCT). 25 bone defects were prepared on sheep mandibles and scanned using a KaVo 3D eXam (KaVo Dental, Biberach, Germany) CBCT device. Section thickness of images were reconstructed at 0.25, 0.5, and 0.75 mm to estimate the volume of these defects using the semiautomatic segmentation method. The volume averages obtained using microcomputed tomography and Archimedes' method served as reference values. The estimated volumes at each section thickness were compared with the actual volumes using the Friedman test. The accuracy of volume estimation was determined by the percentage error with respect to the reference values, and the mean absolute error (MAE) was calculated. Volumetric values of bone defects obtained with CBCT at section thicknesses up to 0.5 mm were compatible with the actual volumes (p > 0.05). The percentage errors at section thicknesses of 0.25, 0.5, and 0.75 mm were -5.4%, -7.3%, and -13.1%, respectively. The mean absolute errors were 13.6 mm3, 15.7 mm3, and 18.2 mm3, respectively. The section thickness values of CBCT images can be increased to a reasonable level to obtain accurate volume estimation results and save time. The semiautomatic segmentation method can be used reliably for volume estimations of bone defects.

Accurate Estimation of Commercial Volume in Tropical Forests

Abstract Accurate estimates of commercial volume in tropical forests are key for the implementation of sustainable forest management plans. Because of the lack of local or generic volumetric equations, most forest managers and forestry services are still using traditional expansion factors (i.e., multiplication of the diameter by a given value) to estimate the volume of commercial tree species in the Amazon. Volumetric models were developed through a unique data set of 1,264 fallen trees fully measured in 150 sample plots located across a broad range of forests in Amapá, Brazil. Forest-specific volumetric models were developed and compared with a generic (i.e., across all forests) model and with published equations developed elsewhere in the Amazon. The generic equation performed well in all forest types and allowed precise predictions. The most efficient sampling design to develop volumetric models consists of measuring approximately 50 trees across four different size classes representing the whole population. The form factors (FF) developed locally generated substantial bias but performed better than the traditional FF (0.7). Overall, our results suggest that it is possible to develop accurate generic models to estimate commercial timber volume, and this study can serve as a guideline for forest managers or scientists interested in calibrating volumetric models in a cost-efficient way.

Increasing Volumetric Prediction Accuracy—An Essential Prerequisite for End-Product Forecasting in Red Pine

Sustainable forest management requires accurate estimates of wood volume. At present, red pine (Pinus resinosa Sol. ex Aiton) is the most widely planted conifer tree species in southern Ontario, Canada. Therefore, inside and outside bark volume equations were developed for red pine trees grown in plantations. One hundred and fifty red pine trees were sampled from 30 even-aged plantations from across Ontario, Canada. Height-diameter pairs along the boles of sampled trees used to calculate stem volumes were obtained from stem analysis. Equations fitted to the data were a combined variable, modified combined variable, and modified form of dimensionally compatible volume equations. These equations were compared for their goodness-of-fit statistics, logical consistency, and predictive accuracy. The goodness-of-fit characteristics for all three equations were comparable for both inside and outside bark volumes. However, the estimated values for the intercept for the modified form of the dimensionally compatible and modified combined variable volume equations were negative and nonsignificant. The combined variable volume equation resulted in logically consistent parameter estimates in the presence of random effects parameters. Therefore, this equation was selected as the inside and outside bark volume equation for red pine trees grown in plantations. A nonlinear mixed-effects modeling approach was applied in fitting the final volume equation that included a weight (power function) to address heteroscedasticity. The equations developed here can be used to calculate inside and outside bark volumes of red pine plantations in boreal forests in Eastern Canada. These equations would require both diameter at breast height (DBH) and total height values in meters.

AdQSM: A New Method for Estimating Above-Ground Biomass from TLS Point Clouds

Forest above-ground biomass (AGB) can be estimated based on light detection and ranging (LiDAR) point clouds. This paper introduces an accurate and detailed quantitative structure model (AdQSM), which can estimate the AGB of large tropical trees. AdQSM is based on the reconstruction of 3D tree models from terrestrial laser scanning (TLS) point clouds. It represents a tree as a set of closed and complete convex polyhedra. We use AdQSM to model 29 trees of various species (total 18 species) scanned by TLS from three study sites (the dense tropical forests of Peru, Indonesia, and Guyana). The destructively sampled tree geometry measurement data is used as reference values to evaluate the accuracy of diameter at breast height (DBH), tree height, tree volume, branch volume, and AGB estimated from AdQSM. After AdQSM reconstructs the structure and volume of each tree, AGB is derived by combining the wood density of the specific tree species from destructive sampling. The AGB estimation from AdQSM and the post-harvest reference measurement data show a satisfying agreement. The coefficient of variation of root mean square error (CV-RMSE) and the concordance correlation coefficient (CCC) are 20.37% and 0.97, respectively. AdQSM provides accurate tree volume estimation, regardless of the characteristics of the tree structure, without major systematic deviations. We compared the accuracy of AdQSM and TreeQSM in modeling the volume of 29 trees. The tree volume from AdQSM is compared with the reference value, and the determination coefficient (R2), relative bias (rBias), and CV-RMSE of tree volume are 0.96, 6.98%, and 22.62%, respectively. The tree volume from TreeQSM is compared with the reference value, and the R2, relative Bias (rBias), and CV-RMSE of tree volume are 0.94, −9.69%, and 23.20%, respectively. The CCCs between the volume estimates based on AdQSM, TreeQSM, and the reference values are 0.97 and 0.96. AdQSM also models the branches in detail. The volume of branches from AdQSM is compared with the destructive measurement reference data. The R2, rBias, and CV-RMSE of the branches volume are 0.97, 12.38%, and 36.86%, respectively. The DBH and height of the harvested trees were used as reference values to test the accuracy of AdQSM’s estimation of DBH and tree height. The R2, rBias, and CV-RMSE of DBH are 0.94, −5.01%, and 9.06%, respectively. The R2, rBias, and CV-RMSE of the tree height were 0.95, 1.88%, and 5.79%, respectively. This paper provides not only a new QSM method for estimating AGB based on TLS point clouds but also the potential for further development and testing of allometric equations.