Species-level classification of urban trees from worldview-2 imagery in Debrecen, Hungary: an effective tool for planning a comprehensive green network to reduce dust pollution

Urban green spaces of cities are crucial elements of city structure that ensure habitat for species and ecological functionality of habitat patches, maintain biodiversity, and provide environmental services. However, detailed maps intended for planning and improving the existing network require a quick and effective technique for assessing the possibilities. Multispectral imagery is an accessible source for species-level classification of urban trees. Using a multispectral image from the WorldView–2 satellite sensor, we classified six of the most common urban tree species in Debrecen, Hungary. Maximum Likelihood (ML) and Support Vector Machine (SVM) classifiers were applied to different numbers of the MNF-transformed bands. The best overall accuracy was achieved with the ML algorithm applied to the first four transformed bands (75.1%), and with the SVM algorithm applied to eight bands (71.0%). In general, ML performed better than SVM. Despite the relatively low number of spectral bands, we achieved moderately good accuracy for basic vegetation mapping, which can be used in spatial planning and decision making. In a future interdisciplinary research study, we could merge the classification results with the dust adsorption capacity of individual species to assess the reduction of dust pollution by urban trees.

Download Full-text

Remote Estimation of Mangrove Aboveground Carbon Stock at the Species Level Using a Low-Cost Unmanned Aerial Vehicle System

Remote Sensing ◽

10.3390/rs11091018 ◽

2019 ◽

Vol 11 (9) ◽

pp. 1018 ◽

Cited By ~ 4

Author(s):

Zhen Li ◽

Qijie Zan ◽

Qiong Yang ◽

Dehuang Zhu ◽

Youjun Chen ◽

...

Keyword(s):

Remote Sensing ◽

Carbon Stock ◽

Low Cost ◽

Important Variable ◽

Species Level ◽

Canopy Height ◽

Individual Species ◽

Support Vector ◽

Mangrove Forests ◽

Aboveground Carbon

There is ongoing interest in developing remote sensing technology to map and monitor the spatial distribution and carbon stock of mangrove forests. Previous research has demonstrated that the relationship between remote sensing derived parameters and aboveground carbon (AGC) stock varies for different species types. However, the coarse spatial resolution of satellite images has restricted the estimated AGC accuracy, especially at the individual species level. Recently, the availability of unmanned aerial vehicles (UAVs) has provided an operationally efficient approach to map the distribution of species and accurately estimate AGC stock at a fine scale in mangrove areas. In this study, we estimated mangrove AGC in the core area of northern Shenzhen Bay, South China, using four kinds of variables, including species type, canopy height metrics, vegetation indices, and texture features, derived from a low-cost UAV system. Three machine-learning algorithm models, including Random Forest (RF), Support Vector Regression (SVR), and Artificial Neural Network (ANN), were compared in this study, where a 10-fold cross-validation was used to evaluate each model’s effectiveness. The results showed that a model that used all four type of variables, which were based on the RF algorithm, provided better AGC estimates (R2 = 0.81, relative RMSE (rRMSE) = 0.20, relative MAE (rMAE) = 0.14). The average predicted AGC from this model was 93.0 ± 24.3 Mg C ha−1, and the total estimated AGC was 7903.2 Mg for the mangrove forests. The species-based model had better performance than the considered canopy-height-based model for AGC estimation, and mangrove species was the most important variable among all the considered input variables; the mean height (Hmean) the second most important variable. Additionally, the RF algorithms showed better performance in terms of mangrove AGC estimation than the SVR and ANN algorithms. Overall, a low-cost UAV system with a digital camera has the potential to enable satisfactory predictions of AGC in areas of homogenous mangrove forests.

Download Full-text

IMPROVING CLASSIFICATION OF MULTISPECTRAL IMAGES BASED ON SELECTED RELATIONS

ISPRS - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences ◽

10.5194/isprs-archives-xliii-b3-2020-377-2020 ◽

2020 ◽

Vol XLIII-B3-2020 ◽

pp. 377-381

Author(s):

L. Cohen ◽

O. Almog ◽

M. Shoshany

Keyword(s):

Land Cover ◽

Spectral Band ◽

Research Area ◽

Support Vector ◽

Land Cover Types ◽

Spectral Bands ◽

Definition Of ◽

Spectral Representations ◽

The City

Abstract. A novel classification technique based on definition of unique spectral relations (such as slopes among spectral bands) for all land cover types named (SSF Significant Spectral Features) is presented in the article.A large slopes combination between spectral band pairs is calculated and spectral characterizations that emphasizes the best spectral land cover separation is sought. Increasing in dimensionality of spectral representations is balanced by the simplicity of calculations. The technique has been examined on data acquired by a flown hyperspectral scanner (AISA). The spectral data was narrowed into the equivalent 8 world-view2 channels. The research area was in the city of “Hadera”, Israel, which included 10 land cover types in an urban area, open area and road infrastructure. The comparison between the developed SSF technique and common techniques such as: SVM (Support Vector Machine) and ML (Maximum Likelihood) has shown a clear advantage over ML technique, while produced similar results as SVM. The poorest results of using SSF technique was achieved in an herbaceous area (70%). However, the simplicity of the method, the well-defined parameters it produces for interpreting the results, makes it intuitive over using techniques such as SVM, which is considered as a not explicit classifier.

Download Full-text

Evaluation of linear discriminant and support vector machine classifiers for classification of nitrogen status in mature oil palm from SPOT-6 satellite images: analysis of raw spectral bands and spectral indices

Geocarto International ◽

10.1080/10106049.2018.1434687 ◽

2018 ◽

Vol 34 (7) ◽

pp. 735-749

Author(s):

Amiratul Diyana Amirruddin ◽

Farrah Melissa Muharam

Keyword(s):

Support Vector Machine ◽

Oil Palm ◽

Satellite Images ◽

Support Vector ◽

Spectral Indices ◽

Nitrogen Status ◽

Linear Discriminant ◽

Spectral Bands

Download Full-text

Spatial Mutual Information Based Hyperspectral Band Selection for Classification

The Scientific World JOURNAL ◽

10.1155/2015/630918 ◽

2015 ◽

Vol 2015 ◽

pp. 1-7 ◽

Cited By ~ 3

Author(s):

Anthony Amankwah

Keyword(s):

Mutual Information ◽

Supervised Classification ◽

Band Selection ◽

Support Vector ◽

Spatial Dependency ◽

Information Redundancy ◽

Spectral Bands ◽

Validation Criteria ◽

Selection For

The amount of information involved in hyperspectral imaging is large. Hyperspectral band selection is a popular method for reducing dimensionality. Several information based measures such as mutual information have been proposed to reduce information redundancy among spectral bands. Unfortunately, mutual information does not take into account the spatial dependency between adjacent pixels in images thus reducing its robustness as a similarity measure. In this paper, we propose a new band selection method based on spatial mutual information. As validation criteria, a supervised classification method using support vector machine (SVM) is used. Experimental results of the classification of hyperspectral datasets show that the proposed method can achieve more accurate results.

Download Full-text

Classification of Hyperspectral Images with CNN in Agricultural Lands

Biology and Life Sciences Forum ◽

10.3390/iecag2021-09739 ◽

2021 ◽

Vol 3 (1) ◽

pp. 6

Author(s):

Eren Can Seyrek ◽

Murat Uysal

Keyword(s):

Classification Performance ◽

Hyperspectral Images ◽

Superior Performance ◽

Support Vector ◽

World Population ◽

Agricultural Lands ◽

Leading Role ◽

Spectral Bands ◽

Land Cover Maps

Hyperspectral images (HSI) offer detailed spectral reflectance information about sensed objects through provision of information on hundreds of narrow spectral bands. HSI have a leading role in a broad range of applications, such as in forestry, agriculture, geology, and environmental sciences. The monitoring and management of agricultural lands is of great importance for meeting the nutritional and other needs of a rapidly and continuously increasing world population. In relation to this, classification of HSI is an effective way for creating land use and land cover maps quickly and accurately. In recent years, classification of HSI using convolutional neural networks (CNN), which is a sub-field of deep learning, has become a very popular research topic and several CNN architectures have been developed by researchers. The aim of this study was to investigate the classification performance of CNN model on agricultural HSI scenes. For this purpose, a 3D-2D CNN framework and a well-known support vector machine (SVM) model were compared using the Indian Pines and Salinas Scene datasets that contain crop and mixed vegetation classes. As a result of this study, it was confirmed that use of 3D-2D CNN offers superior performance for classifying agricultural HSI datasets.

Download Full-text

COMPARISON OF LANDSAT-8 AND SENTINEL-2 DATA FOR CLASSIFICATION OF RABI CROPS OVER KARNATAKA, INDIA

ISPRS - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences ◽

10.5194/isprs-archives-xlii-3-w6-579-2019 ◽

2019 ◽

Vol XLII-3/W6 ◽

pp. 579-584

Author(s):

S. Paul ◽

D. N. Kumar

Keyword(s):

Vegetation Index ◽

Support Vector ◽

Growth Stages ◽

Landsat 8 ◽

Large Spatial Scale ◽

Spectral Bands ◽

Sentinel 2A ◽

Rabi Crops ◽

Sentinel 2

Abstract. Classification of crops is very important to study different growth stages and forecast yield. Remote sensing data plays a significant role in crop identification and condition assessment over a large spatial scale. Importance of Normalized Difference Indices (NDIs) along with surface reflectances of remotely sensed spectral bands have been evaluated for classification of eight types of Rabi crops utilizing the Landsat-8 and Sentinel-2 datasets and performances of both the satellites are compared. Landsat-8 and Sentinel-2A images are acquired for the location of crops and seven and nine spectral bands are utilized respectively for the classification. Experiments are carried out considering the different combinations of surface reflectances of spectral bands and optimal NDIs as features in support vector machine classifier. Optimal NDIs are selected from the set of 7C2 and 9C2 NDIs of Landsat-8 and Sentinel-2A datasets respectively using the partial informational correlation measure, a nonparametric feature selection approach. Few important vegetation indices (e.g. enhanced vegetation index) are also experimented in combination with the surface reflectances and NDIs to perform the crop classification. It has been observed that combination of surface reflectances and optimal NDIs can classify the crops more efficiently. The average overall accuracy of 80.96% and 88.16% are achieved using the Landsat-8 and Sentinel-2A datasets respectively. It has been observed that all the crop classes except Paddy and Cotton achieve producer accuracy and user accuracy of more than 75% and 85% respectively. This technique can be implemented for crop identification with adequate accessibility of crop information.

Download Full-text

Foliar venation characteristics of Vaccinium sect. Macropelma and sect. Myrtillus (Ericaceae): a comparative analysis

Canadian Journal of Botany ◽

10.1139/b97-913 ◽

1997 ◽

Vol 75 (11) ◽

pp. 2015-2025 ◽

Cited By ~ 2

Author(s):

E. Ann Powell ◽

S. P. Vander Kloet

Keyword(s):

Cluster Analysis ◽

Comparative Analysis ◽

Diagnostic Value ◽

Bundle Sheath ◽

Species Level ◽

Individual Species ◽

Diagnostic Ability ◽

Operational Taxonomic Units ◽

Hawaiian Species

A suite of eight foliar venation characteristics from two closely related sections of Vaccinium L., sect. Macropelma Klotzsch and sect. Myrtillus Dumortier, were analyzed to assess the level of similarity between these sister sections, address the questionable classification of Vaccinium cereum (L.f) Forster, and assess the diagnostic ability of venation characteristics. Our analysis revealed no significant differences in foliar venation characteristics between V. cereum and the Hawaiian species from sect. Macropelma. In addition, all the leaves of V. cereum examined contained the attenuated extended bundle sheath system as found in other representatives from sect. Macropelma. Taximetric analysis showed that these quantitative foliar venation characters were diagnostic at the sectional level but not at the species level. A cluster analysis of 108 operational taxonomic units did separate the two sections but did not discriminate among the 12 individual species. However, a similar taximetric analysis using the character means for each species (for the eight foliar characters), which "smoothed out" the variation due to environmental plasticity, produced a dendrogram which closely resembled the currently accepted Vaccinium classification. Although further investigations into the constancy and reliability of foliar venation characters in the remaining sections of Vaccinium are desirable, our results clearly show that these quantitative vegetative foliar characters do have diagnostic value and can be utilized in taximetric analyses. Key words: Vaccinium, venation, leaf, systematics.

Download Full-text

DETECTION OF SHALLOW WATER AREA WITH MACHINE LEARNING ALGORITHMS

ISPRS - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences ◽

10.5194/isprs-archives-xlii-2-w13-1269-2019 ◽

2019 ◽

Vol XLII-2/W13 ◽

pp. 1269-1273 ◽

Cited By ~ 3

Author(s):

N. Yagmur ◽

N. Musaoglu ◽

G. Taskin

Keyword(s):

Support Vector Machine ◽

Shallow Water ◽

Spatial Resolution ◽

Water Area ◽

Water Bodies ◽

Support Vector ◽

Spectral Bands ◽

Shallow Water Area ◽

Sentinel 2

Abstract. Remote sensing techniques has been widely used for detecting water bodies in especially wetlands. Different classification methods and water indices has used for this purpose and there are numerous studies for detecting water bodies. However, detecting shallow water area is difficult comparing with deep water bodies because of the mixed pixels. Akgol Wetland is chosen as study area to detect shallow water. For this purpose, Sentinel 2 satellite image, which gives more accurate results thanks to higher spatial resolution than the images having medium spatial resolution, is used. In this study, two classification approaches were applied on Sentinel 2 image to detect shallow water area. In the first approach, effectiveness of indices was determined and classification of spectral bands with indices shows higher accuracy than classification of only spectral bands by using support vector machine classification method. In the second approach, support vector machine recursive feature elimination method used for the most effective features in the first approach. Besides overall accuracy of only spectral bands is obtained as 88.10%, spectral bands and indices’ accuracy was obtained as 91.84%.

Download Full-text