Luming Fang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Luming Fang has a total of 21 co-authors. Here are the top co-authors by number of shared publications:
- Yuhui Weng
Arthur Temple College of Forestry and Agriculture, Stephen F. Austin State University, Nacogdoches, USA
3 shared publications - Gengsheng Fang
College of Mathematics and Computer Science, Zhejiang A & F University, Hangzhou 311300, China
2 shared publications - Xiaobing He
Baishanzu Scientific Research Monitoring Center, Qianjiangyuan-Baishanzu National Park, Lishui 323000, China
2 shared publications - Xinyu Zheng
College of Mathematics and Computer Science, Zhejiang A&F University, Hangzhou 311300, China
2 shared publications - Chongzhi Chen
College of Environmental and Resource Sciences, Zhejiang University, Hangzhou 310058, China
1 shared publication - Clinton Talmage
Arthur Temple College of Forestry and Agriculture, Stephen F. Austin State University, Nacogdoches, TX 75965, USA
1 shared publication
Latest Publications
- Open Access
Development of a Real-Time Continuous Measurement System for Tree Radial Direction
Tree diameter at breast height (DBH) is the most fundamental factor in modelling tree growth, but current DBH measurement instruments mainly focus on instantaneous acquisition, making it difficult to measure tree growth continuously and accurately. In this study, we propose a wireless sensing network that can transmit data in signal-free environments, and combine sensor and computer technologies to develop a real-time continuous measurement system for tree radials, which has the advantages of working in real-time, being low-cost and stable, and enabling high-precision. It can be applied to the DBH measurement of trees in the range of 50 mm–380 mm, with a measurement accuracy of 0.001 mm. Additionally, whole-point sampling, conducted 24 h per day, integrates DBH data measurement, transmission, storage and visualization analysis. After measuring in the field for a year, it initially reveals the change in DBH within the test area within that year. This study provides a scientific basis for researching the microscopic growth pattern of trees and establishing a tree growth model, which will be further optimised and improved in terms of appearance structure, communication and power supply in the future.
Open AccessDevelopment of a Real-Time Continuous Measurement System for Tree Radial Direction
Tree diameter at breast height (DBH) is the most fundamental factor in modelling tree growth, but current DBH measurement instruments mainly focus on instantaneous acquisition, making it difficult to measure tree growth continuously and accurately. In this study, we propose a wireless sensing network that can transmit data in signal-free environments, and combine sensor and computer technologies to develop a real-time continuous measurement system for tree radials, which has the advantages of working in real-time, being low-cost and stable, and enabling high-precision. It can be applied to the DBH measurement of trees in the range of 50 mm–380 mm, with a measurement accuracy of 0.001 mm. Additionally, whole-point sampling, conducted 24 h per day, integrates DBH data measurement, transmission, storage and visualization analysis. After measuring in the field for a year, it initially reveals the change in DBH within the test area within that year. This study provides a scientific basis for researching the microscopic growth pattern of trees and establishing a tree growth model, which will be further optimised and improved in terms of appearance structure, communication and power supply in the future.
- Open Access
Synergistic Use of Sentinel-1 and Sentinel-2 Based on Different Preprocessing for Predicting Forest Aboveground Biomass
Forest aboveground biomass (AGB, Mg/ha) measurement is one of the key indicators for carbon storage evaluation. Remote sensing techniques have been widely employed to predict forest AGB. However, little attention has been paid to the implications involved in the preprocessing of satellite data. In this work, considering the areas of low forest AGB in our survey plots, we explored the implications of employing atmospheric correction and speckle filtering with Sentinel-1 (S1) synthetic aperture radar (SAR) and Sentinel-2 (S2) to predict forest AGB using multiple linear regression (MLR) and extreme gradient boosting (XGBoost). In the present study, the types of plots examined included oaks (Quercus spp.), Chinese firs (Cunninghamia lanceolata), and Masson pines (Pinus massoniana), and all of the plots were investigated. Specifically, the feature variables related to S1 (dual polarization and texture measures) and S2 (spectral bands) were modeled individually, and 16 feature sets, including different combinations of S1 and S2 based on different preprocessing measures, were established using MLR and XGBoost. The results show that speckle filtering and atmospheric correction marginally influenced the capacity of the S2 spectral bands, the SAR dual-polarization backscatter, and the SAR-based textural measures in predicting the AGB in our survey plots. The associations between the speckle-filtered and unfiltered SAR images and the S2 Top-of-Atmosphere and Bottom-of-Atmosphere products were considerably strong. Additionally, the texture models generally showed better performances than the raw SAR data. Ultimately, the groups that only encompassed the S2 spectral bands were the best-performing groups among the 16 feature sets, while the groups that included only S1-based data generally performed the worst.
Open AccessSynergistic Use of Sentinel-1 and Sentinel-2 Based on Different Preprocessing for Predicting Forest Aboveground Biomass
Forest aboveground biomass (AGB, Mg/ha) measurement is one of the key indicators for carbon storage evaluation. Remote sensing techniques have been widely employed to predict forest AGB. However, little attention has been paid to the implications involved in the preprocessing of satellite data. In this work, considering the areas of low forest AGB in our survey plots, we explored the implications of employing atmospheric correction and speckle filtering with Sentinel-1 (S1) synthetic aperture radar (SAR) and Sentinel-2 (S2) to predict forest AGB using multiple linear regression (MLR) and extreme gradient boosting (XGBoost). In the present study, the types of plots examined included oaks (Quercus spp.), Chinese firs (Cunninghamia lanceolata), and Masson pines (Pinus massoniana), and all of the plots were investigated. Specifically, the feature variables related to S1 (dual polarization and texture measures) and S2 (spectral bands) were modeled individually, and 16 feature sets, including different combinations of S1 and S2 based on different preprocessing measures, were established using MLR and XGBoost. The results show that speckle filtering and atmospheric correction marginally influenced the capacity of the S2 spectral bands, the SAR dual-polarization backscatter, and the SAR-based textural measures in predicting the AGB in our survey plots. The associations between the speckle-filtered and unfiltered SAR images and the S2 Top-of-Atmosphere and Bottom-of-Atmosphere products were considerably strong. Additionally, the texture models generally showed better performances than the raw SAR data. Ultimately, the groups that only encompassed the S2 spectral bands were the best-performing groups among the 16 feature sets, while the groups that included only S1-based data generally performed the worst.
- Open Access
- 31 May 2023
- Published by MDPI AG in Sustainability
- Vol. 15 (11) , 8888
Assessing Forest Quality through Forest Growth Potential, an Index Based on Improved CatBoost Machine Learning
Human activities have always depended on nature, and forests are an important part of this; the determination and improvement of forest quality is therefore highly significant. Currently, domestic and foreign research on forest quality focuses on the current states of forests. We propose a new research direction based on the future states. By referencing and analyzing the forest quality standards of domestic and foreign experts and institutions, the concept and model for calculating forest growth potential were constructed. Forest growth potential is a new forest quality indicator. Based on the data of 110,000 subcompartments of forest resources from the Lin’an and Landsat8 satellites’ remote sensing data, the unit volume was predicted using three machine-learning algorithms: random gradient descent SGD, the integrated machine learning algorithm CatBoost, and deep learning CNN. The CatBoost algorithm model was improved based on Optuna; then the improved CatBoost algorithm was selected through evaluation indicators for the prediction of forest volume and finally incorporated into the calculation model for forest growth-potential value. The forest growth-potential value was calculated, and an accurate forest quality improvement scheme based on the subcompartments is preliminarily discussed. The successful calculation of forest growth potential values has a certain reference significance, providing guidance for accurately improving forest quality and forest management. The improved CatBoost calculation model is effective in the prediction of forest growth potential, and the determination coefficient R2 reaches 0.89, a value that compares favorably with those in other studies.
- 31 May 2023
- Published by MDPI AG in Sustainability
- Vol. 15 (11) , 8888
Open AccessAssessing Forest Quality through Forest Growth Potential, an Index Based on Improved CatBoost Machine Learning
Human activities have always depended on nature, and forests are an important part of this; the determination and improvement of forest quality is therefore highly significant. Currently, domestic and foreign research on forest quality focuses on the current states of forests. We propose a new research direction based on the future states. By referencing and analyzing the forest quality standards of domestic and foreign experts and institutions, the concept and model for calculating forest growth potential were constructed. Forest growth potential is a new forest quality indicator. Based on the data of 110,000 subcompartments of forest resources from the Lin’an and Landsat8 satellites’ remote sensing data, the unit volume was predicted using three machine-learning algorithms: random gradient descent SGD, the integrated machine learning algorithm CatBoost, and deep learning CNN. The CatBoost algorithm model was improved based on Optuna; then the improved CatBoost algorithm was selected through evaluation indicators for the prediction of forest volume and finally incorporated into the calculation model for forest growth-potential value. The forest growth-potential value was calculated, and an accurate forest quality improvement scheme based on the subcompartments is preliminarily discussed. The successful calculation of forest growth potential values has a certain reference significance, providing guidance for accurately improving forest quality and forest management. The improved CatBoost calculation model is effective in the prediction of forest growth potential, and the determination coefficient R2 reaches 0.89, a value that compares favorably with those in other studies.
- Open Access
- 29 May 2023
- Published by MDPI AG in Remote Sensing
- Vol. 15 (11) , 2821
Texture Features Derived from Sentinel-2 Vegetation Indices for Estimating and Mapping Forest Growing Stock Volume
Forest growing stock volume (GSV, m3/ha) is an indispensable variable for forest resources supervision. Precise measurement of GSV is conducive to monitoring forest dynamics. Nevertheless, little research has explored the pattern pixel values and the function of texture features derived from a vegetation index (VI) for GSV prediction. In this study, we investigated combining linear regression or Random Forest with Sentinel-2 spectral predictors and image textures derived from spectral bands and vegetation indices, which were based on standard deviation values or mean values at the plot level, for predicting GSV of Masson pine, theropencedrymion and all the survey plots in Anji County, China. Specifically, ten groups of experiments encompassing combinations of spectral parameters and texture measures were established for detecting the potential of image textures based on two different image pixel statistics. The results showed that texture measures derived from VI were superior to spectral parameters or texture measures derived from spectral bands for estimating the GSV of single tree species. Moreover, texture features based on shortwave infrared bands or their related VI and the standard deviation at pixel level for spectral bands/indices were emphasized. Finally, the mean value at the plot level exhibited slightly stronger potential than the standard deviation in general.
- 29 May 2023
- Published by MDPI AG in Remote Sensing
- Vol. 15 (11) , 2821
Open AccessTexture Features Derived from Sentinel-2 Vegetation Indices for Estimating and Mapping Forest Growing Stock Volume
Forest growing stock volume (GSV, m3/ha) is an indispensable variable for forest resources supervision. Precise measurement of GSV is conducive to monitoring forest dynamics. Nevertheless, little research has explored the pattern pixel values and the function of texture features derived from a vegetation index (VI) for GSV prediction. In this study, we investigated combining linear regression or Random Forest with Sentinel-2 spectral predictors and image textures derived from spectral bands and vegetation indices, which were based on standard deviation values or mean values at the plot level, for predicting GSV of Masson pine, theropencedrymion and all the survey plots in Anji County, China. Specifically, ten groups of experiments encompassing combinations of spectral parameters and texture measures were established for detecting the potential of image textures based on two different image pixel statistics. The results showed that texture measures derived from VI were superior to spectral parameters or texture measures derived from spectral bands for estimating the GSV of single tree species. Moreover, texture features based on shortwave infrared bands or their related VI and the standard deviation at pixel level for spectral bands/indices were emphasized. Finally, the mean value at the plot level exhibited slightly stronger potential than the standard deviation in general.
- Open Access
- 21 September 2022
- Published by MDPI AG in Remote Sensing
- Vol. 14 (19) , 4702
Landsat Data Based Prediction of Loblolly Pine Plantation Attributes in Western Gulf Region, USA
The suitability of using Landsat sensor variables to predict key stand attributes, including stand average dominant/codominant tree height (HT), mean diameter at breast height (DBH), the number of trees per hectare (NT), basal area per hectare (BA), and stand density index (SDI), of intensively managed loblolly pine plantations in the Western Gulf Region at the plot/stand level was assessed. In total, thirty Landsat sensor variables including six original bands, three vegetation indices, three Tasseled Cap transformed indices, and eighteen texture measure variables were used as predictors. Field data of 125 permanent plots located across east Texas and western Louisiana were used as reference data. Individual trees of those plots were measured at plot establishment (referred to as the first cycle measurement; average about 4.5 years old) and remeasured in three-year intervals (the second cycle measurement at approximately seven years old and the third cycle measurement at approximately 10 years old). Thus, field reference data represent stand development from open- (first cycle) to closed-canopy (third cycle). Models to predict stand HT, DBH, NT, BA, and SDI were developed by cycle using multiple linear regression (MLR) and also random forests (RF) methods. Results indicated that the first cycle stands HT, DBH, BA, and SDI were well predicted using the Landsat sensor variables with R2 > 0.7 and low RMSEs. These relationships weakened with stand age, although still moderate with R2 being around 0.45 for the second cycle measurement and became practically useless (R2 < 0.30) for the third cycle measurement. For NT, no meaningful models were achieved regardless of the measurement cycle. The MLR and RF models were comparable in accuracy and had similar key predictors. Overall, the shortwave infrared bands, red band, and wetness index were the most important predictors, but their dominance declined with the cycle. Texture measure variables were relatively less important but a trend of increasing their importance with cycle was noted. Results show promise for operationally predicting stand variables for young pine plantations, an age class that typically presents significant challenges using conventional forest measurement methodologies. Potential methods to further improve model accuracy and how to use the results within the context of pine plantation management planning in the region were discussed.
- 21 September 2022
- Published by MDPI AG in Remote Sensing
- Vol. 14 (19) , 4702
Open AccessLandsat Data Based Prediction of Loblolly Pine Plantation Attributes in Western Gulf Region, USA
The suitability of using Landsat sensor variables to predict key stand attributes, including stand average dominant/codominant tree height (HT), mean diameter at breast height (DBH), the number of trees per hectare (NT), basal area per hectare (BA), and stand density index (SDI), of intensively managed loblolly pine plantations in the Western Gulf Region at the plot/stand level was assessed. In total, thirty Landsat sensor variables including six original bands, three vegetation indices, three Tasseled Cap transformed indices, and eighteen texture measure variables were used as predictors. Field data of 125 permanent plots located across east Texas and western Louisiana were used as reference data. Individual trees of those plots were measured at plot establishment (referred to as the first cycle measurement; average about 4.5 years old) and remeasured in three-year intervals (the second cycle measurement at approximately seven years old and the third cycle measurement at approximately 10 years old). Thus, field reference data represent stand development from open- (first cycle) to closed-canopy (third cycle). Models to predict stand HT, DBH, NT, BA, and SDI were developed by cycle using multiple linear regression (MLR) and also random forests (RF) methods. Results indicated that the first cycle stands HT, DBH, BA, and SDI were well predicted using the Landsat sensor variables with R2 > 0.7 and low RMSEs. These relationships weakened with stand age, although still moderate with R2 being around 0.45 for the second cycle measurement and became practically useless (R2 < 0.30) for the third cycle measurement. For NT, no meaningful models were achieved regardless of the measurement cycle. The MLR and RF models were comparable in accuracy and had similar key predictors. Overall, the shortwave infrared bands, red band, and wetness index were the most important predictors, but their dominance declined with the cycle. Texture measure variables were relatively less important but a trend of increasing their importance with cycle was noted. Results show promise for operationally predicting stand variables for young pine plantations, an age class that typically presents significant challenges using conventional forest measurement methodologies. Potential methods to further improve model accuracy and how to use the results within the context of pine plantation management planning in the region were discussed.
- Open Access
- 1 January 2021
- Published by Wiley in Journal of Sensors
- Vol. 2021 (1)
A Method to Locate Tree Positions Using Ultrawideband Technology
Tree position plays an important role in research on forest resources and ecological functions, and quickly and accurately obtaining tree position data has long been the focus of investigators. However, the classical method is time‐consuming and laborious; thus, a convenient method of measuring tree position is needed. The primary achievements of this study include the following: (1) a device was designed for precise location of trees; (2) a new location algorithm was proposed for pentagonal localization based on the received signal strength indication and ultrawideband technology; and (3) a PC software application was developed for automatically storing and uploading tree position data. The device was applied to 10 circular plots with a diameter of 24 m to test the positioning speed and accuracy. The results showed that the tree positions could be accurately estimated. On the x‐ and y‐axes, the biases were ‐3.94 and 3.36 cm, respectively, and the root mean square errors (RMSEs) were 28.39 and 28.53 cm, respectively. The mean error (Ed) between the estimated and reference distances was 36.13 cm, and the standard deviation was 16.67 cm. The device is inexpensive and easy to use and carry in the field; thus, it is suitable for locating trees in environments with complex terrain.
- 1 January 2021
- Published by Wiley in Journal of Sensors
- Vol. 2021 (1)
Open AccessA Method to Locate Tree Positions Using Ultrawideband Technology
Tree position plays an important role in research on forest resources and ecological functions, and quickly and accurately obtaining tree position data has long been the focus of investigators. However, the classical method is time‐consuming and laborious; thus, a convenient method of measuring tree position is needed. The primary achievements of this study include the following: (1) a device was designed for precise location of trees; (2) a new location algorithm was proposed for pentagonal localization based on the received signal strength indication and ultrawideband technology; and (3) a PC software application was developed for automatically storing and uploading tree position data. The device was applied to 10 circular plots with a diameter of 24 m to test the positioning speed and accuracy. The results showed that the tree positions could be accurately estimated. On the x‐ and y‐axes, the biases were ‐3.94 and 3.36 cm, respectively, and the root mean square errors (RMSEs) were 28.39 and 28.53 cm, respectively. The mean error (Ed) between the estimated and reference distances was 36.13 cm, and the standard deviation was 16.67 cm. The device is inexpensive and easy to use and carry in the field; thus, it is suitable for locating trees in environments with complex terrain.
- Open Access
An Integrated Method for Coding Trees, Measuring Tree Diameter, and Estimating Tree Positions
Accurately measuring tree diameter at breast height (DBH) and estimating tree positions in a sample plot are important in tree mensuration. The main aims of this paper include (1) developing a new, integrated device that can identify trees using the quick response (QR) code technique to record tree identifications, measure DBH, and estimate tree positions concurrently; (2) designing an innovative algorithm to measure DBH using only two angle sensors, which is simple and can reduce the impact of eccentric stems on DBH measures; and (3) designing an algorithm to estimate the position of the tree by combining ultra-wide band (UWB) technology and altitude sensors, which is based on the received signal strength indication (RSSI) algorithm and quadrilateral localization algorithm. This novel device was applied to measure ten 10 × 10 m square plots of diversified environments and various tree species to test its accuracy. Before measuring a plot, a coded sticker was fixed at a height of 1.3 m on each individual tree stem, and four UWB module anchors were set up at the four corners of the plot. All individual trees’ DBHs and positions within the plot were then measured. Tree DBH, measured using a tree caliper, and the values of tree positions, measured using tape, angle ruler, and inclinometer, were used as the respective reference values for comparison. Across the plots, the decode rate of QR codes was 100%, with an average response time less than two seconds. The DBH values had a bias of 1.89 mm (1.88% in relative terms) and a root mean square error (RMSE) of 5.38 mm (4.53% in relative terms). The tree positions were accurately estimated; the biases on the x-axis and the y-axis of the tree position were −8.55–14.88 cm and −12.07–24.49 cm, respectively, and the corresponding RMSEs were 12.94–33.96 cm and 17.78–28.43 cm. The average error between the estimated and reference distances was 30.06 cm, with a standard deviation of 13.53 cm. The device is cheap and friendly to use in addition to its high accuracy. Although further studies are needed, our method provides a great alternative to conventional tools for improving the efficiency and accuracy of tree mensuration.
Open AccessAn Integrated Method for Coding Trees, Measuring Tree Diameter, and Estimating Tree Positions
Accurately measuring tree diameter at breast height (DBH) and estimating tree positions in a sample plot are important in tree mensuration. The main aims of this paper include (1) developing a new, integrated device that can identify trees using the quick response (QR) code technique to record tree identifications, measure DBH, and estimate tree positions concurrently; (2) designing an innovative algorithm to measure DBH using only two angle sensors, which is simple and can reduce the impact of eccentric stems on DBH measures; and (3) designing an algorithm to estimate the position of the tree by combining ultra-wide band (UWB) technology and altitude sensors, which is based on the received signal strength indication (RSSI) algorithm and quadrilateral localization algorithm. This novel device was applied to measure ten 10 × 10 m square plots of diversified environments and various tree species to test its accuracy. Before measuring a plot, a coded sticker was fixed at a height of 1.3 m on each individual tree stem, and four UWB module anchors were set up at the four corners of the plot. All individual trees’ DBHs and positions within the plot were then measured. Tree DBH, measured using a tree caliper, and the values of tree positions, measured using tape, angle ruler, and inclinometer, were used as the respective reference values for comparison. Across the plots, the decode rate of QR codes was 100%, with an average response time less than two seconds. The DBH values had a bias of 1.89 mm (1.88% in relative terms) and a root mean square error (RMSE) of 5.38 mm (4.53% in relative terms). The tree positions were accurately estimated; the biases on the x-axis and the y-axis of the tree position were −8.55–14.88 cm and −12.07–24.49 cm, respectively, and the corresponding RMSEs were 12.94–33.96 cm and 17.78–28.43 cm. The average error between the estimated and reference distances was 30.06 cm, with a standard deviation of 13.53 cm. The device is cheap and friendly to use in addition to its high accuracy. Although further studies are needed, our method provides a great alternative to conventional tools for improving the efficiency and accuracy of tree mensuration.