scholarly journals Tree Species Mapping on Sentinel-2 Satellite Imagery with Weakly Supervised Classification and Object-Wise Sampling

Forests ◽  
2021 ◽  
Vol 12 (10) ◽  
pp. 1413
Author(s):  
Svetlana Illarionova ◽  
Alexey Trekin ◽  
Vladimir Ignatiev ◽  
Ivan Oseledets
Keyword(s):  
Spatial Distribution ◽  
Satellite Imagery ◽  
Tree Species ◽  
Dominant Species ◽  
Specific Information ◽  
Data Set ◽  
Weakly Supervised ◽  
Markup Adjustment ◽  
Tree Type ◽  
Sentinel 2

Information on forest composition, specifically tree types and their distribution, aids in timber stock calculation and can help to better understand the biodiversity in a particular region. Automatic satellite imagery analysis can significantly accelerate the process of tree type classification, which is traditionally carried out by ground-based observation. Although computer vision methods have proven their efficiency in remote sensing tasks, specific challenges arise in forestry applications. The forest inventory data often contain the tree type composition but do not describe their spatial distribution within each individual stand. Therefore, some pixels can be assigned a wrong label in the semantic segmentation task if we consider each stand to be homogeneously populated by its dominant species. Another challenge is the spatial distribution of individual stands within the study area. Classes are usually imbalanced and distributed nonuniformly that makes sampling choice more critical. This study aims to enhance tree species classification based on a neural network approach providing automatic markup adjustment and improving sampling technique. For forest species markup adjustment, we propose using a weakly supervised learning approach based on the knowledge of dominant species content within each stand. We also propose substituting the commonly used CNN sampling approach with the object-wise one to reduce the effect of the spatial distribution of forest stands. We consider four species commonly found in Russian boreal forests: birch, aspen, pine, and spruce. We use imagery from the Sentinel-2 satellite, which has multiple bands (in the visible and infrared spectra) and a spatial resolution of up to 10 meters. A data set of images for Leningrad Oblast of Russia is used to assess the methods. We demonstrate how to modify the training strategy to outperform a basic CNN approach from F1-score 0.68 to 0.76. This approach is promising for future studies to obtain more specific information about stands composition even using incomplete data.

scholarly journals Classifying Tree Species in Sentinel-2 Satellite Imagery Using Weakly Supervised Neural Networks

2020 ◽  
Author(s):  
Svetlana Illarionova ◽  
Alexey Trekin ◽  
Vladimir Ignatiev ◽  
Ivan Oseledets
Keyword(s):  
Neural Networks ◽  
Satellite Imagery ◽  
Tree Species ◽  
Weakly Supervised ◽  
Sentinel 2

scholarly journals IMPROVING LULC CLASSIFICATION FROM SATELLITE IMAGERY USING DEEP LEARNING – EUROSAT DATASET

2021 ◽  
Vol XLIII-B3-2021 ◽  
pp. 369-376
Author(s):  
H. Yassine ◽  
K. Tout ◽  
M. Jaber
Keyword(s):  
Machine Learning ◽  
Deep Learning ◽  
Satellite Imagery ◽  
Vegetation Index ◽  
Near Infrared ◽  
Spectral Feature ◽  
Classification Performance ◽  
Training Data ◽  
Data Set ◽  
Sentinel 2

Abstract. Machine learning (ML) has proven useful for a very large number of applications in several domains. It has realized a remarkable growth in remote-sensing image analysis over the past few years. Deep Learning (DL) a subset of machine learning were applied in this work to achieve a better classification of Land Use Land Cover (LULC) in satellite imagery using Convolutional Neural Networks (CNNs). EuroSAT benchmarking data set is used as training data set which uses Sentinel-2 satellite images. Sentinel-2 provides images with 13 spectral feature bands, but surprisingly little attention has been paid to these features in deep learning models. The majority of applications focused only on using RGB due to high availability of the RGB models in computer vision. While RGB gives an accuracy of 96.83% using CNN, we are presenting two approaches to improve the classification performance of Sentinel-2 images. In the first approach, features are extracted from 13 spectral feature bands of Sentinel-2 instead of RGB which leads to accuracy of 98.78%. In the second approach features are extracted from 13 spectral bands of Sentinel-2 in addition to calculated indices used in LULC like Blue Ratio (BR), Vegetation index based on Red Edge (VIRE) and Normalized Near Infrared (NNIR), etc. which gives a better accuracy of 99.58%.

scholarly journals Potential of High-Resolution Satellite Imagery for Mapping Distribution and Evaluating Ecological Characteristics of Tree Species at the Angkor Monument, Cambodia

2015 ◽  
Vol 7 (3) ◽  
pp. 49-59
Author(s):  
Mizuki Tomita ◽  
Yoshihiko Hirabuki ◽  
Yuji Araki ◽  
Shinji Tsukawaki ◽  
Bora Ly ◽  
...  
Keyword(s):  
High Resolution ◽  
Satellite Imagery ◽  
Tree Species ◽  
Dominant Species ◽  
Forest Composition ◽  
Individual Species ◽  
Clear Understanding ◽  
Ecological Traits ◽  
Study Plot ◽  
Large Trees

Abstract Large trees play several vital roles in the Angkor monuments landscape. They protect biodiversity, enhance the tourism experience, and provide various ecosystem services to local residents. A clear understanding of forest composition and distribution of individual species, as well as timely monitoring of changes, is necessary for conservation of these trees. using traditional field work, obtaining this sort of data is time-consuming and labour-intensive. This research investigates classification of very high resolution remote sensing data as a tool for efficient analyses. QuickBird satellite imagery was used to clarify the tree species community in and around Preah Khan temple, to elucidate differences in ecological traits among the three dominant species (Dipterocarpus alatus, Lagerstroemia calyculata and Tetrameles nudiflora), and to identify crowns of the dominant species. Population structures of trees were determined using a 14.26ha study plot. Species name, DBH, height, height under the crown and crown area were recorded for all trees over 40 cm in DBH. Tree locations were also recorded so as to provide references for the imagery analysis. Ecological traits of the dominant species were estimated using regressions by an expanded allometric equation for both large and small trees, based on DBH, height, height under the crown and crown width. The total number of species in the study plot was 45. From a spatial perspective, the three dominant species over 100 cm in DBH were segregated from each other. D. alatus, L. calyculata and T. nudiflora were concentrated, respectively, along the approach to the temple, near the centre of the complex, and on the walls of the monument. Object Based Image Analysis (OBIA) conducted using QuickBird satellite imagery, showed that crowns of D. alatus were largely determined by maximum of NIR layer and mean of digital number in panchromatic layer. Differences in the parameters for both asymptotic height and spatial distribution among the dominant species, result from differences in ecological traits, and enhance the value of the tourism resource by providing a dramatic shift of forest scenery that can be enjoyed by visitors to the monument.

scholarly journals Tree Species Classification with Multi-Temporal Sentinel-2 Data

2018 ◽  
Vol 10 (11) ◽  
pp. 1794 ◽  
Author(s):  
Magnus Persson ◽  
Eva Lindberg ◽  
Heather Reese
Keyword(s):  
Variable Selection ◽  
Tree Species ◽  
Pedunculate Oak ◽  
Species Classification ◽  
Data Set ◽  
Tree Species Classification ◽  
Red Edge ◽  
Lower Accuracy ◽  
Multi Temporal ◽  
Sentinel 2

The Sentinel-2 program provides the opportunity to monitor terrestrial ecosystems with a high temporal and spectral resolution. In this study, a multi-temporal Sentinel-2 data set was used to classify common tree species over a mature forest in central Sweden. The tree species to be classified were Norway spruce (Picea abies), Scots pine (Pinus silvestris), Hybrid larch (Larix × marschlinsii), Birch (Betula sp.) and Pedunculate oak (Quercus robur). Four Sentinel-2 images from spring (7 April and 27 May), summer (9 July) and fall (19 October) of 2017 were used along with the Random Forest (RF) classifier. A variable selection approach was implemented to find fewer and uncorrelated bands resulting in the best model for tree species identification. The final model resulting in the highest overall accuracy (88.2%) came from using all bands from the four image dates. The single image that gave the most accurate classification result (80.5%) was the late spring image (27 May); the 27 May image was always included in subsequent image combinations that gave the highest overall accuracy. The five tree species were classified with a user’s accuracy ranging from 70.9% to 95.6%. Thirteen of the 40 bands were selected in a variable selection procedure and resulted in a model with only slightly lower accuracy (86.3%) than that using all bands. Among the highest ranked bands were the red edge bands 2 and 3 as well as the narrow NIR (near-infrared) band 8a, all from the 27 May image, and SWIR (short-wave infrared) bands from all four image dates. This study shows that the red-edge bands and SWIR bands from Sentinel-2 are of importance, and confirms that spring and/or fall images capturing phenological differences between the species are most useful to tree species classification.

scholarly journals Hyperdominance in Amazonian forest carbon cycling

2015 ◽  
Vol 6 (1) ◽  
Author(s):  
Sophie Fauset ◽  
Michelle O. Johnson ◽  
Manuel Gloor ◽  
Timothy R. Baker ◽  
Abel Monteagudo M. ◽  
...  
Keyword(s):  
Carbon Cycling ◽  
Tree Species ◽  
Dominant Species ◽  
Maximum Size ◽  
Abundant Species ◽  
Data Set ◽  
Forest Function ◽  
Amazonian Forests ◽  
Global Carbon ◽  
Amazonian Trees

Abstract While Amazonian forests are extraordinarily diverse, the abundance of trees is skewed strongly towards relatively few ‘hyperdominant’ species. In addition to their diversity, Amazonian trees are a key component of the global carbon cycle, assimilating and storing more carbon than any other ecosystem on Earth. Here we ask, using a unique data set of 530 forest plots, if the functions of storing and producing woody carbon are concentrated in a small number of tree species, whether the most abundant species also dominate carbon cycling, and whether dominant species are characterized by specific functional traits. We find that dominance of forest function is even more concentrated in a few species than is dominance of tree abundance, with only ≈1% of Amazon tree species responsible for 50% of carbon storage and productivity. Although those species that contribute most to biomass and productivity are often abundant, species maximum size is also influential, while the identity and ranking of dominant species varies by function and by region.

scholarly journals A Deep Learning Approach to an Enhanced Building Footprint and Road Detection in High-Resolution Satellite Imagery

2021 ◽  
Vol 13 (16) ◽  
pp. 3135
Author(s):  
Christian Ayala ◽  
Rubén Sesma ◽  
Carlos Aranda ◽  
Mikel Galar
Keyword(s):  
Deep Learning ◽  
High Resolution ◽  
Spatial Resolution ◽  
Satellite Imagery ◽  
Input Image ◽  
Road Detection ◽  
Data Set ◽  
Building Footprint ◽  
High Resolution Satellite Imagery ◽  
Sentinel 2

The detection of building footprints and road networks has many useful applications including the monitoring of urban development, real-time navigation, etc. Taking into account that a great deal of human attention is required by these remote sensing tasks, a lot of effort has been made to automate them. However, the vast majority of the approaches rely on very high-resolution satellite imagery (<2.5 m) whose costs are not yet affordable for maintaining up-to-date maps. Working with the limited spatial resolution provided by high-resolution satellite imagery such as Sentinel-1 and Sentinel-2 (10 m) makes it hard to detect buildings and roads, since these labels may coexist within the same pixel. This paper focuses on this problem and presents a novel methodology capable of detecting building and roads with sub-pixel width by increasing the resolution of the output masks. This methodology consists of fusing Sentinel-1 and Sentinel-2 data (at 10 m) together with OpenStreetMap to train deep learning models for building and road detection at 2.5 m. This becomes possible thanks to the usage of OpenStreetMap vector data, which can be rasterized to any desired resolution. Accordingly, a few simple yet effective modifications of the U-Net architecture are proposed to not only semantically segment the input image, but also to learn how to enhance the resolution of the output masks. As a result, generated mappings quadruplicate the input spatial resolution, closing the gap between satellite and aerial imagery for building and road detection. To properly evaluate the generalization capabilities of the proposed methodology, a data-set composed of 44 cities across the Spanish territory have been considered and divided into training and testing cities. Both quantitative and qualitative results show that high-resolution satellite imagery can be used for sub-pixel width building and road detection following the proper methodology.

scholarly journals Calling for improved quality in the registration of traditional Chinese medicine during the public health emergency: a survey of trial registries for COVID-19, H1N1, and SARS

Trials ◽  
2021 ◽  
Vol 22 (1) ◽  
Author(s):  
Zhuoran Kuang ◽  
◽  
Xiaoyan Li ◽  
Jianxiong Cai ◽  
Yaolong Chen ◽  
...  
Keyword(s):  
Clinical Trial ◽  
Clinical Trials ◽  
Chinese Medicine ◽  
Traditional Chinese Medicine ◽  
Secondary Outcome ◽  
Specific Information ◽  
Clinical Trial Registry ◽  
Data Set ◽  
Diagnosis Criteria

Abstract Objective To assess the registration quality of traditional Chinese medicine (TCM) clinical trials for COVID-19, H1N1, and SARS. Method We searched for clinical trial registrations of TCM in the WHO International Clinical Trials Registry Platform (ICTRP) and Chinese Clinical Trial Registry (ChiCTR) on April 30, 2020. The registration quality assessment is based on the WHO Trial Registration Data Set (Version 1.3.1) and extra items for TCM information, including TCM background, theoretical origin, specific diagnosis criteria, description of intervention, and outcomes. Results A total of 136 records were examined, including 129 severe acute respiratory syndrome coronavirus 2 (COVID-19) and 7 H1N1 influenza (H1N1) patients. The deficiencies in the registration of TCM clinical trials (CTs) mainly focus on a low percentage reporting detailed information about interventions (46.6%), primary outcome(s) (37.7%), and key secondary outcome(s) (18.4%) and a lack of summary result (0%). For the TCM items, none of the clinical trial registrations reported the TCM background and rationale; only 6.6% provided the TCM diagnosis criteria or a description of the TCM intervention; and 27.9% provided TCM outcome(s). Conclusion Overall, although the number of registrations of TCM CTs increased, the registration quality was low. The registration quality of TCM CTs should be improved by more detailed reporting of interventions and outcomes, TCM-specific information, and sharing of the result data.

scholarly journals Do galactic bars depend on environment?: an information theoretic analysis of Galaxy Zoo 2

2020 ◽  
Vol 501 (1) ◽  
pp. 994-1001
Author(s):  
Suman Sarkar ◽  
Biswajit Pandey ◽  
Snehasish Bhattacharjee
Keyword(s):  
Spatial Distribution ◽  
Mutual Information ◽  
Local Density ◽  
Statistical Significance ◽  
Distribution Functions ◽  
Cumulative Distribution ◽  
Host Galaxy ◽  
Data Sets ◽  
Data Set ◽  
Information Theoretic

ABSTRACT We use an information theoretic framework to analyse data from the Galaxy Zoo 2 project and study if there are any statistically significant correlations between the presence of bars in spiral galaxies and their environment. We measure the mutual information between the barredness of galaxies and their environments in a volume limited sample (Mr ≤ −21) and compare it with the same in data sets where (i) the bar/unbar classifications are randomized and (ii) the spatial distribution of galaxies are shuffled on different length scales. We assess the statistical significance of the differences in the mutual information using a t-test and find that both randomization of morphological classifications and shuffling of spatial distribution do not alter the mutual information in a statistically significant way. The non-zero mutual information between the barredness and environment arises due to the finite and discrete nature of the data set that can be entirely explained by mock Poisson distributions. We also separately compare the cumulative distribution functions of the barred and unbarred galaxies as a function of their local density. Using a Kolmogorov–Smirnov test, we find that the null hypothesis cannot be rejected even at $75{{\ \rm per\ cent}}$ confidence level. Our analysis indicates that environments do not play a significant role in the formation of a bar, which is largely determined by the internal processes of the host galaxy.

scholarly journals Soil carbon and nutrient stocks under Scots pine plantations in comparison to European beech forests: a paired-plot study across forests with different management history and precipitation regimes

2021 ◽  
Vol 8 (1) ◽  
Author(s):  
Marco Diers ◽  
Robert Weigel ◽  
Heike Culmsee ◽  
Christoph Leuschner
Keyword(s):  
Scots Pine ◽  
Tree Species ◽  
European Beech ◽  
Organic Layer ◽  
Species Identity ◽  
Data Set ◽  
Nutrient Pools ◽  
Total N ◽  
Precipitation Regimes ◽  
Pine Stands

Abstract Background Organic carbon stored in forest soils (SOC) represents an important element of the global C cycle. It is thought that the C storage capacity of the stable pool can be enhanced by increasing forest productivity, but empirical evidence in support of this assumption from forests differing in tree species and productivity, while stocking on similar substrate, is scarce. Methods We determined the stocks of SOC and macro-nutrients (nitrogen, phosphorus, calcium, potassium and magnesium) in nine paired European beech/Scots pine stands on similar Pleistocene sandy substrates across a precipitation gradient (560–820 mm∙yr− 1) in northern Germany and explored the influence of tree species, forest history, climate, and soil pH on SOC and nutrient pools. Results While the organic layer stored on average about 80% more C under pine than beech, the pools of SOC and total N in the total profile (organic layer plus mineral soil measured to 60 cm and extrapolated to 100 cm) were greater under pine by about 40% and 20%, respectively. This contrasts with a higher annual production of foliar litter and a much higher fine root biomass in beech stands, indicating that soil C sequestration is unrelated to the production of leaf litter and fine roots in these stands on Pleistocene sandy soils. The pools of available P and basic cations tended to be higher under beech. Neither precipitation nor temperature influenced the SOC pool, whereas tree species was a key driver. An extended data set (which included additional pine stands established more recently on former agricultural soil) revealed that, besides tree species identity, forest continuity is an important factor determining the SOC and nutrient pools of these stands. Conclusion We conclude that tree species identity can exert a considerable influence on the stocks of SOC and macronutrients, which may be unrelated to productivity but closely linked to species-specific forest management histories, thus masking weaker climate and soil chemistry effects on pool sizes.

Sign in / Sign up

Export Citation Format

Share Document