scholarly journals URBAN MATERIAL CLASSIFICATION USING SPECTRAL AND TEXTURAL FEATURES RETRIEVED FROM AUTOENCODERS

2020 ◽  
Vol V-1-2020 ◽  
pp. 25-32
Author(s):  
R. Ilehag ◽  
J. Leitloff ◽  
M. Weinmann ◽  
A. Schenk
Keyword(s):  
Classification Tree ◽  
Remote Sensing Data ◽  
Hyperspectral Data ◽  
Gradient Boosting ◽  
Support Vector ◽  
Textural Features ◽  
Spectral Features ◽  
Material Classification ◽  
Characteristic Features ◽  
Spectral Libraries

Abstract. Classification of urban materials using remote sensing data, in particular hyperspectral data, is common practice. Spectral libraries can be utilized to train a classifier since they provide spectral features about selected urban materials. However, urban materials can have similar spectral characteristic features due to high inter-class correlation which can lead to misclassification. Spectral libraries rarely provide imagery of their samples, which disables the possibility of classifying urban materials with additional textural information. Thus, this paper conducts material classification comparing the benefits of using close-range acquired spectral and textural features. The spectral features consist of either the original spectra, a PCA-based encoding or the compressed spectral representation of the original spectra retrieved using a deep autoencoder. The textural features are generated using a deep denoising convolutional autoencoder. The spectral and textural features are gathered from the recently published spectral library KLUM. Three classifiers are used, the two well-established Random Forest and Support Vector Machine classifiers in addition to a Histogram-based Gradient Boosting Classification Tree. The achieved overall accuracy was within the range of 70–80% with a standard deviation between 2–10% across all classification approaches. This indicates that the amount of samples still is insufficient for some of the material classes for this classification task. Nonetheless, the classification results indicate that the spectral features are more important for assigning material labels than the textural features.

scholarly journals Early Detection of Plant Viral Disease Using Hyperspectral Imaging and Deep Learning

Sensors ◽  
2021 ◽  
Vol 21 (3) ◽  
pp. 742
Author(s):  
Canh Nguyen ◽  
Vasit Sagan ◽  
Matthew Maimaitiyiming ◽  
Maitiniyazi Maimaitijiang ◽  
Sourav Bhadra ◽  
...  
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Early Detection ◽  
Convolutional Neural Network ◽  
Near Infrared ◽  
Hyperspectral Data ◽  
Viral Diseases ◽  
Support Vector ◽  
Spectral Features ◽  
Feature Spaces

Early detection of grapevine viral diseases is critical for early interventions in order to prevent the disease from spreading to the entire vineyard. Hyperspectral remote sensing can potentially detect and quantify viral diseases in a nondestructive manner. This study utilized hyperspectral imagery at the plant level to identify and classify grapevines inoculated with the newly discovered DNA virus grapevine vein-clearing virus (GVCV) at the early asymptomatic stages. An experiment was set up at a test site at South Farm Research Center, Columbia, MO, USA (38.92 N, −92.28 W), with two grapevine groups, namely healthy and GVCV-infected, while other conditions were controlled. Images of each vine were captured by a SPECIM IQ 400–1000 nm hyperspectral sensor (Oulu, Finland). Hyperspectral images were calibrated and preprocessed to retain only grapevine pixels. A statistical approach was employed to discriminate two reflectance spectra patterns between healthy and GVCV vines. Disease-centric vegetation indices (VIs) were established and explored in terms of their importance to the classification power. Pixel-wise (spectral features) classification was performed in parallel with image-wise (joint spatial–spectral features) classification within a framework involving deep learning architectures and traditional machine learning. The results showed that: (1) the discriminative wavelength regions included the 900–940 nm range in the near-infrared (NIR) region in vines 30 days after sowing (DAS) and the entire visual (VIS) region of 400–700 nm in vines 90 DAS; (2) the normalized pheophytization index (NPQI), fluorescence ratio index 1 (FRI1), plant senescence reflectance index (PSRI), anthocyanin index (AntGitelson), and water stress and canopy temperature (WSCT) measures were the most discriminative indices; (3) the support vector machine (SVM) was effective in VI-wise classification with smaller feature spaces, while the RF classifier performed better in pixel-wise and image-wise classification with larger feature spaces; and (4) the automated 3D convolutional neural network (3D-CNN) feature extractor provided promising results over the 2D convolutional neural network (2D-CNN) in learning features from hyperspectral data cubes with a limited number of samples.

scholarly journals Evaluating Variable Selection and Machine Learning Algorithms for Estimating Forest Heights by Combining Lidar and Hyperspectral Data

2020 ◽  
Vol 9 (9) ◽  
pp. 507
Author(s):  
Sanjiwana Arjasakusuma ◽  
Sandiaga Swahyu Kusuma ◽  
Stuart Phinn
Keyword(s):  
Machine Learning ◽  
Feature Selection ◽  
Learning Algorithms ◽  
Principal Component ◽  
Hyperspectral Data ◽  
Machine Learning Algorithms ◽  
Gradient Boosting ◽  
Support Vector ◽  
Forest Height ◽  
Extreme Gradient Boosting

Machine learning has been employed for various mapping and modeling tasks using input variables from different sources of remote sensing data. For feature selection involving high- spatial and spectral dimensionality data, various methods have been developed and incorporated into the machine learning framework to ensure an efficient and optimal computational process. This research aims to assess the accuracy of various feature selection and machine learning methods for estimating forest height using AISA (airborne imaging spectrometer for applications) hyperspectral bands (479 bands) and airborne light detection and ranging (lidar) height metrics (36 metrics), alone and combined. Feature selection and dimensionality reduction using Boruta (BO), principal component analysis (PCA), simulated annealing (SA), and genetic algorithm (GA) in combination with machine learning algorithms such as multivariate adaptive regression spline (MARS), extra trees (ET), support vector regression (SVR) with radial basis function, and extreme gradient boosting (XGB) with trees (XGbtree and XGBdart) and linear (XGBlin) classifiers were evaluated. The results demonstrated that the combinations of BO-XGBdart and BO-SVR delivered the best model performance for estimating tropical forest height by combining lidar and hyperspectral data, with R2 = 0.53 and RMSE = 1.7 m (18.4% of nRMSE and 0.046 m of bias) for BO-XGBdart and R2 = 0.51 and RMSE = 1.8 m (15.8% of nRMSE and −0.244 m of bias) for BO-SVR. Our study also demonstrated the effectiveness of BO for variables selection; it could reduce 95% of the data to select the 29 most important variables from the initial 516 variables from lidar metrics and hyperspectral data.

scholarly journals Seabed Mapping in Coastal Shallow Waters Using High Resolution Multispectral and Hyperspectral Imagery

2018 ◽  
Vol 10 (8) ◽  
pp. 1208 ◽  
Author(s):  
Javier Marcello ◽  
Francisco Eugenio ◽  
Javier Martín ◽  
Ferran Marqués
Keyword(s):  
High Resolution ◽  
Aquatic Vegetation ◽  
Hyperspectral Imagery ◽  
Remote Sensing Data ◽  
Hyperspectral Data ◽  
Support Vector ◽  
Shallow Waters ◽  
Benthic Habitats ◽  
Spectral Angle Mapper ◽  
Seabed Mapping

Coastal ecosystems experience multiple anthropogenic and climate change pressures. To monitor the variability of the benthic habitats in shallow waters, the implementation of effective strategies is required to support coastal planning. In this context, high-resolution remote sensing data can be of fundamental importance to generate precise seabed maps in coastal shallow water areas. In this work, satellite and airborne multispectral and hyperspectral imagery were used to map benthic habitats in a complex ecosystem. In it, submerged green aquatic vegetation meadows have low density, are located at depths up to 20 m, and the sea surface is regularly affected by persistent local winds. A robust mapping methodology has been identified after a comprehensive analysis of different corrections, feature extraction, and classification approaches. In particular, atmospheric, sunglint, and water column corrections were tested. In addition, to increase the mapping accuracy, we assessed the use of derived information from rotation transforms, texture parameters, and abundance maps produced by linear unmixing algorithms. Finally, maximum likelihood (ML), spectral angle mapper (SAM), and support vector machine (SVM) classification algorithms were considered at the pixel and object levels. In summary, a complete processing methodology was implemented, and results demonstrate the better performance of SVM but the higher robustness of ML to the nature of information and the number of bands considered. Hyperspectral data increases the overall accuracy with respect to the multispectral bands (4.7% for ML and 9.5% for SVM) but the inclusion of additional features, in general, did not significantly improve the seabed map quality.

scholarly journals Multispectral Approach for Identifying Invasive Plant Species Based on Flowering Phenology Characteristics

2019 ◽  
Vol 11 (8) ◽  
pp. 953 ◽  
Author(s):  
Tarin Paz-Kagan ◽  
Micha Silver ◽  
Natalya Panov ◽  
Arnon Karnieli
Keyword(s):  
Plant Species ◽  
Species Distribution ◽  
Invasive Plant ◽  
Remote Sensing Data ◽  
Hyperspectral Data ◽  
Support Vector ◽  
Temperature Pattern ◽  
Multispectral Images ◽  
Invasive Plant Species ◽  
Distribution Maps

Invasive plant species (IPS) are the second biggest threat to biodiversity after habitat loss. Since the spatial extent of IPS is essential for managing the invaded ecosystem, the current study aims at identifying and mapping the aggressive IPS of Acacia salicina and Acacia saligna, to understand better the key factors influencing their distribution in the coastal plain of Israel. This goal was achieved by integrating airborne-derived hyperspectral imaging and multispectral earth observation for creating species distribution maps. Hyperspectral data, in conjunction with high spatial resolution species distribution maps, were used to train the multispectral images at the species level. We incorporated a series of statistical models to classify the IPS location and to recognize their distribution and density. We took advantage of the phenological flowering stages of Acacia trees, as obtained by the multispectral images, for the support vector machine classification procedure. The classification yielded an overall Kappa coefficient accuracy of 0.89. We studied the effect of various environmental and human factors on IPS density by using a random forest machine learning model, to understand the mechanisms underlying successful invasions, and to assess where IPS have a higher likelihood of occurring. This algorithm revealed that the high density of Acacia most closely related to elevation, temperature pattern, and distances from rivers, settlements, and roads. Our results demonstrate how the integration of remote-sensing data with different data sources can assist in determining IPS proliferation and provide detailed geographic information for conservation and management efforts to prevent their future spread.

scholarly journals Lithostratigraphic Classification Method Combining Optimal Texture Window Size Selection and Test Sample Purification Using Landsat 8 OLI Data

2018 ◽  
Vol 10 (1) ◽  
pp. 565-581 ◽  
Author(s):  
Yufang Qiu ◽  
Dongping Ming
Keyword(s):  
Classification Accuracy ◽  
Window Size ◽  
Test Sample ◽  
Classification Method ◽  
Support Vector ◽  
Textural Features ◽  
Spectral Features ◽  
Size Selection ◽  
Sample Purification ◽  
Texture Extraction

Abstract Gray Level Co-Occurrence Matrix (GLCM), as a measure of spatial features has been used as supplemental information to improve image classification accuracy for lithological recognition. Window size is an important parameter for texture extraction, which will affect the extracted texture results. Besides, the existence of mixed pixels in image usually causes errors in test samples, which significantly influences the credibility of accuracy assessment. Thus, this paper proposes a lithological classification method combined with optimal texture window size selection and test sample purification. Firstly, optimal window size pre-estimated based on semivariogram was used to calculated GLCM texture of image. Secondly, based on multidimensional textural and spectral features, a support vector machine (SVM) classifier was employed to classify the image. Thirdly, using the proposed sample purification method and textural features of image, sample purification rules were created based on attribute coherence to remove the test sample points that conflicted with the rules. Finally, the validity of the semivariogram-based texture extraction window selection was verified by classifications based on Angular Second Moment (ASM) of different window sizes combined with spectral features. Also, the accuracies between different combinations of classifications were assessed by test samples with and without sample purification. Experimental results show that the pre-estimated texture window size can guarantee a classification result with high classification accuracy for lithological classification. The results also demonstrated that the accuracy of lithological classification based on spectral features and ASM textural features was the highest. The overall lithological classification accuracy and kappa value, without sample purification selected by stratified sampling, were respectively 87.4% and 0.84, however those with sample purification were respectively 88.01% and 0.85. The results show that the proposed method is capable of yielding more reliable lithostratigraphic identification.

Research on Classification of Remote Sensing Image Based on SVM Including Textural Features

2014 ◽  
Vol 543-547 ◽  
pp. 2559-2565 ◽  
Author(s):  
Feng Hua Huang
Keyword(s):  
Remote Sensing ◽  
Land Use ◽  
Support Vector Machine ◽  
Remote Sensing Image ◽  
Classification Method ◽  
Likelihood Method ◽  
Support Vector ◽  
Textural Features ◽  
Spectral Features ◽  
Experimental Area

In order to solve the problems in the traditional remote sensing image based on spectral information, such as low classification accuracy, different object with the same spectral features or the same object with the different spectral features, and limited sample quantity and so on, a remote sensing image classification method based on the support vector machine (SVM) including with textural features is proposed. Using Langqi Island of Fuzhou as experimental area, preprocessing and principal component analysis were made to initialize TM images, and the spectral features and GLCM-based textural features of ground objects were extracted and analyzed respectively. Then, the extraction, training and testing of samples based on the two types of features were finished for training various SVM classifiers, which were used for classifying land use in the experimental area. Through the maximum likelihood method, the BP neural network and the support vector machine (SVM), a crossed classification and contrast experiment was made to two different types of samples based on the simple spectral features and the features combined with texture respectively. The experimental results showed that the SVM classification method including textural features can effectively improve the accuracy of land use classification, and therefore it can be promoted better.

scholarly journals Predicting Asteroid Types: Importance of Individual and Combined Features

2021 ◽  
Vol 8 ◽  
Author(s):  
Hanna Klimczak ◽  
Wojciech Kotłowski ◽  
Dagmara Oszkiewicz ◽  
Francesca DeMeo ◽  
Agnieszka Kryszczyńska ◽  
...  
Keyword(s):  
Adaptive Learning ◽  
Surface Composition ◽  
High Rate ◽  
Stochastic Gradient Descent ◽  
Batch Size ◽  
Gradient Boosting ◽  
Support Vector ◽  
Spectral Features ◽  
Spectral Slope ◽  
And Performance

Asteroid taxonomies provide a link to surface composition and mineralogy of those objects, although that connection is not fully unique. Currently, one of the most commonly used asteroid taxonomies is that of Bus-DeMeo. The spectral range covering 0.45–2.45 μm is used to assign a taxonomic class in that scheme. Such observations are only available for a few hundreds of asteroids (out of over one million). On the other hand, a growing amount of space and ground-based surveys delivers multi-filter photometry, which is often used in predicting asteroid types. Those surveys are typically dedicated to studying other astronomical objects, and thus not optimized for asteroid taxonomic classifications. The goal of this study was to quantify the importance and performance of different asteroid spectral features, parameterizations, and methods in predicting the asteroid types. Furthermore, we aimed to identify the key spectral features that can be used to optimize future surveys toward asteroid characterization. Those broad surveys typically are restricted to a few bands; therefore, selecting those that best link them to asteroid taxonomy is crucial in light of maximizing the science output for solar system studies. First, we verified that with the increased number of asteroid spectra, the Bus–DeMeo procedure to create taxonomy still produces the same overall scheme. Second, we confirmed that machine learning methods such as naive Bayes, support vector machine (SVM), gradient boosting, and multilayer networks can reproduce that taxonomic classification at a high rate of over 81% balanced accuracy for types and 93% for complexes. We found that multilayer perceptron with three layers of 32 neurons and stochastic gradient descent solver, batch size of 32, and adaptive learning performed the best in the classification task. Furthermore, the top five features (spectral slope and reflectance at 1.05, 0.9, 0.65, and 1.1 μm) are enough to obtain a balanced accuracy of 93% for the prediction of complexes and six features (spectral slope and reflectance at 1.4, 1.05, 0.9, 0.95, and 0.65 μm) to obtain 81% balanced accuracy for taxonomic types. Thus, to optimize future surveys toward asteroid classification, we recommend using filters that cover those features.

scholarly journals Evaluation of Deep Learning Convolutional Neural Network for Crop Classification

2019 ◽  
Vol 8 (2) ◽  
pp. 3960-3963
Keyword(s):  
Neural Network ◽  
Remote Sensing ◽  
Deep Learning ◽  
Random Forest ◽  
Convolutional Neural Network ◽  
Crop Production ◽  
Remote Sensing Data ◽  
Hyperspectral Data ◽  
Support Vector ◽  
Random Forest Classification

In this paper, we have done exploratory experiments using deep learning convolutional neural network framework to classify crops into cotton, sugarcane and mulberry. In this contribution we have used Earth Observing-1 hyperion hyperspectral remote sensing data as the input. Structured data has been extracted from hyperspectral data using a remote sensing tool. An analytical assessment shows that convolutional neural network (CNN) gives more accuracy over classical support vector machine (SVM) and random forest methods. It has been observed that accuracy of SVM is 75 %, accuracy of random forest classification is 78 % and accuracy of CNN using Adam optimizer is 99.3 % and loss is 2.74 %. CNN using RMSProp also gives the same accuracy 99.3 % and the loss is 4.43 %. This identified crop information will be used for finding crop production and for deciding market strategies

scholarly journals Improved Classification of Mangroves Health Status Using Hyperspectral Remote Sensing Data

2014 ◽  
Vol XL-8 ◽  
pp. 667-670 ◽  
Author(s):  
R. Vidhya ◽  
D. Vijayasekaran ◽  
M. Ahamed Farook ◽  
S. Jai ◽  
M. Rohini ◽  
...  
Keyword(s):  
Remote Sensing ◽  
Health Status ◽  
Remote Sensing Data ◽  
Principal Component ◽  
Hyperspectral Remote Sensing ◽  
Hyperspectral Data ◽  
Support Vector ◽  
Spectral Angle Mapper ◽  
Svm Classification ◽  
Sensing Data

Mangrove ecosystem plays a crucial role in costal conservation and provides livelihood supports to humans. It is seriously affected by the various climatic and anthropogenic induced changes. The continuous monitoring is imperative to protect this fragile ecosystem. In this study, the mangrove area and health status has been extracted from Hyperspectral remote sensing data (EO- 1Hyperion) using support vector machine classification (SVM). The principal component transformation (PCT) technique is used to perform the band reduction in Hyperspectral data. The soil adjusted vegetation Indices (SAVI) were used as additional parameters. The mangroves are classified into three classes degraded, healthy and sparse. The SVM classification is generated overall accuracy of 73 % and kappa of 0.62. The classification results were compared with the results of spectral angle mapper classification (SAM). The SAVI also included in SVM classification and the accuracy found to be improved to 82 %. The sparse and degraded mangrove classes were well separated. The results indicate that the mapping of mangrove health is accurate when the machine learning classifier like SVM combined with different indices derived from hyperspectral remote sensing data.

scholarly journals A Unified Framework for Dimensionality Reduction and Classification of Hyperspectral Data

2014 ◽  
Vol XL-8 ◽  
pp. 447-453 ◽  
Author(s):  
P. Kolluru ◽  
K. Pandey ◽  
H. Padalia
Keyword(s):  
Dimensionality Reduction ◽  
Classification Accuracy ◽  
Mineral Exploration ◽  
Remote Sensing Data ◽  
Single Step ◽  
Hyperspectral Data ◽  
Support Vector ◽  
Unified Framework ◽  
Intrinsic Dimensionality

The processing of hyperspectral remote sensing data, for information retrieval, is challenging due to its higher dimensionality. Machine learning based algorithms such as Support Vector Machine (SVM) is preferably applied to perform classification of high dimensionality data. A single-step unified framework is required which could decide the intrinsic dimensionality of data and achieve higher classification accuracy using SVM. This work present development of a SVM-based dimensionality reduction and classification (SVMDRC) framework for hyperspectral data. The proposed unified framework was tested at Los Tollos in Rodalquilar district of Spain, which have predominance of alunite, kaolinite, and illite minerals with sparse vegetation cover. Summer season image was utilized for implementing the proposed method. Modified broken stick rule (MBSR) was used to calculate the intrinsic dimensionality of HyMap data which automatically reduce the number of bands. Comparison of SVMDRC with SVM clearly suggests that SVM alone is inadequate in yielding better classification accuracies for minerals from hyperspectral data rather requires dimensionality reduction. Incorporation of modified broken stick method in SVMDRC framework positively influenced the feature separability and provided better classification accuracy. The mineral distribution map produced for the study area would be useful for refining the areas for mineral exploration.

Sign in / Sign up

Export Citation Format

Share Document