• Korean Journal of Remote Sensing
• /
• v.35 no.6_1
• /
• pp.959-971
• /
• 2019

CNN (Convolutional Neural Network) is one representative deep learning algorithm, which can extract high-level spatial and spectral features, and has been applied for hyperspectral image classification. However, one significant drawback behind the application of CNNs in hyperspectral images is the high dimensionality of the data, which increases the training time and processing complexity. To address this problem, several CNN based hyperspectral image classification studies have exploited PCA (Principal Component Analysis) for dimensionality reduction. One limitation to this is that the spectral information of the original image can be lost through PCA. Although it is clear that the use of PCA affects the accuracy and the CNN training time, the impact of PCA for CNN based hyperspectral image classification has been understudied. The purpose of this study is to analyze the quantitative effect of PCA in CNN for hyperspectral image classification. The hyperspectral images were first transformed through PCA and applied into the CNN model by varying the size of the reduced dimensionality. In addition, 2D-CNN and 3D-CNN frameworks were applied to analyze the sensitivity of the PCA with respect to the convolution kernel in the model. Experimental results were evaluated based on classification accuracy, learning time, variance ratio, and training process. The size of the reduced dimensionality was the most efficient when the explained variance ratio recorded 99.7%~99.8%. Since the 3D kernel had higher classification accuracy in the original-CNN than the PCA-CNN in comparison to the 2D-CNN, the results revealed that the dimensionality reduction was relatively less effective in 3D kernel.

• Korean Journal of Remote Sensing
• /
• v.24 no.2
• /
• pp.213-222
• /
• 2008

Although hyperspectral sensing data have shown great potential to derive various surface information that is not usually available from conventional multispectral image, the acquisition of proper hyperspectral image data are often limited. To use ground-based hyperspectral camera image for remote sensing studies, radiometric calibration should be prerequisite. The objective of this study is to develop radiometric calibration procedure to convert image digital number (DN) value to surface reflectance for the 120 bands ground-based hyperspectral camera. Hyperspectral image and spectral measurements were simultaneously obtained from the experimental target that includes 22 different surface materials of diverse spectral characteristics at wavelength range between 400 to 900 nm. Calibration coefficients to convert image DN value to at-sensor radiance were initially derived from the regression equations between the sample image and spectral measurements using ASD spectroradiometer. Assuming that there is no atmospheric effects when the image acquisition and spectral measurements were made at very close distance in ground, we were also able to derive calibration coefficients that directly transform DN value to surface reflectance. However, these coefficients for deriving reflectance values should not be applied when the camera is used for aerial image that contains significant effect from atmosphere and further atmospheric correction procedure is required in such case.

• Proceedings of the KSRS Conference
• /
• 2007.10a
• /
• pp.6-10
• /
• 2007

Classification of hyperspectral images is challenging. A very high dimensional input space requires an exponentially large amount of data to adequately and reliably represent the classes in that space. In other words in order to obtain statistically reliable classification results, the number of necessary training samples increases exponentially as the number of spectral bands increases. However, in many situations, acquisition of the large number of training samples for these high-dimensional datasets may not be so easy. This problem can be overcome by using multiple classifiers. In this paper we compared the effectiveness of two approaches for creating multiple classifiers, feature selection and feature extraction. The methods are based on generating multiple feature subsets by running feature selection or feature extraction algorithm several times, each time for discrimination of one of the classes from the rest. A maximum likelihood classifier is applied on each of the obtained feature subsets and finally a combination scheme was used to combine the outputs of individual classifiers. Experimental results show the effectiveness of feature extraction algorithm for generating multiple classifiers.

• Journal of Korean Society for Geospatial Information Science
• /
• v.25 no.1
• /
• pp.9-17
• /
• 2017

Hyperspectral imagery is used in the land cover classification with the principle component analysis and minimum noise fraction to reduce the data dimensionality and noise. Recently, studies on the supervised classification using various features having spectral information and spatial characteristic have been carried out. In this study, principle component bands and normalized difference vegetation index(NDVI) was utilized in the supervised classification for the land cover classification. To utilize additional information not included in the principle component bands by the hyperspectral imagery, we tried to increase the classification accuracy by using the NDVI. In addition, the extended attribute profiles(EAP) generated using the morphological filter was used as the input data. The random forest algorithm, which is one of the representative supervised classification, was used. The classification accuracy according to the application of various features based on EAP was compared. Two areas was selected in the experiments, and the quantitative evaluation was performed by using reference data. The classification accuracy of the proposed algorithm showed the highest classification accuracy of 85.72% and 91.14% compared with existing algorithms. Further research will need to develop a supervised classification algorithm and additional input datasets to improve the accuracy of land cover classification using hyperspectral imagery.

• Korean Journal of Remote Sensing
• /
• v.32 no.5
• /
• pp.403-412
• /
• 2016

Compact Reconnaissance Imaging Spectrometer for Mars(CRISM) is 489-band hyperspectral camera of Mars Reconnaissance Orbiter(MRO) that provided data used on many mineral researches over Martian surface. For the detection of minerals in planet, mineral index using a few spectral bands have been used. In this study, we applied Matched Filter and Adaptive Cosine Estimator(ACE) target detection algorithm on CRISM data over Gusev Crater: landing site of Spirit(Mars Exploration Rover A) to investigate its mineral distribution. As a result, olivine, pyroxene, magnetite, etc. is detected at Gusev Crater's Columbia Hills. These results are corresponding to the Spirit rover's field survey result. It is expected that hyperspectral target detection algorithms can be used as effective and easy to use method for the detection and mapping of surface minerals in planet.

• International Journal of Computer Science & Network Security
• /
• v.22 no.4
• /
• pp.193-202
• /
• 2022

The classification of hyperspectral imagery (HSI) is essential in the surface of earth observation. Due to the continuous large number of bands, HSI data provide rich information about the object of study; however, it suffers from the curse of dimensionality. Dimensionality reduction is an essential aspect of Machine learning classification. The algorithms based on feature extraction can overcome the data dimensionality issue, thereby allowing the classifiers to utilize comprehensive models to reduce computational costs. This paper assesses and compares two HSI classification techniques. The first is based on the Joint Spatial-Spectral Stacked Autoencoder (JSSSA) method, the second is based on a shallow Artificial Neural Network (SNN), and the third is used the SVM model. The performance of the JSSSA technique is better than the SNN classification technique based on the overall accuracy and Kappa coefficient values. We observed that the JSSSA based method surpasses the SNN technique with an overall accuracy of 96.13% and Kappa coefficient value of 0.95. SNN also achieved a good accuracy of 92.40% and a Kappa coefficient value of 0.90, and SVM achieved an accuracy of 82.87%. The current study suggests that both JSSSA and SNN based techniques prove to be efficient methods for hyperspectral classification of snow features. This work classified the labeled/ground-truth datasets of snow in multiple classes. The labeled/ground-truth data can be valuable for applying deep neural networks such as CNN, hybrid CNN, RNN for glaciology, and snow-related hazard applications.

• Journal of Biosystems Engineering
• /
• v.40 no.4
• /
• pp.417-427
• /
• 2015

Purpose: This study investigated different band selection methods to classify spectrally similar data - obtained from aerial images of healthy citrus canopies and citrus greening disease (Huanglongbing or HLB) infected canopies - using small differences without unmixing endmember components and therefore without the need for an endmember library. However, large number of hyperspectral bands has high redundancy which had to be reduced through band selection. The objective, therefore, was to first select the best set of bands and then detect citrus Huanglongbing infected canopies using these bands in aerial hyperspectral images. Methods: The forward feature selection algorithm (FFSA) was chosen for band selection. The selected bands were used for identifying HLB infected pixels using various classifiers such as K nearest neighbor (KNN), support vector machine (SVM), naïve Bayesian classifier (NBC), and generalized local discriminant bases (LDB). All bands were also utilized to compare results. Results: It was determined that a few well-chosen bands yielded much better results than when all bands were chosen, and brought the classification results on par with standard hyperspectral classification techniques such as spectral angle mapper (SAM) and mixture tuned matched filtering (MTMF). Median detection accuracies ranged from 66-80%, which showed great potential toward rapid detection of the disease. Conclusions: Among the methods investigated, a support vector machine classifier combined with the forward feature selection algorithm yielded the best results.

• Proceedings of the KSRS Conference
• /
• 2003.11a
• /
• pp.1076-1080
• /
• 2003

$CO_2$ uptake of vegetation is one of the important variables in order to estimate photosynthetic activity, plant growth and carbon budget estimations. The objective of this research was to develop a new estimation method of $CO_2$ uptake of vegetation based on airborne hyperspectral remote sensing measurements in combination with a photosynthetic rate curve model. In this study, a compact airborne spectrographic imager (CASI) was used to obtain image over a field that had been set up to study the $CO_2$ uptake of corn on August 7, 2002. Also, a field survey was conducted concurrently with the CASI overpass. As a field survey, chlorophyll a content, photosynthetic rate curve, Leaf area, dry biomass and light condition were measured. The developed estimation method for $CO_2$ uptake consists of three major parts: a linear mixture model, an enhanced big leaf model and a photosynthetic rate curve model. The Accuracy of this scheme indicates that $CO_2$ uptake of vegetation could be estimated by using airborne hyperspectral remote sensing data in combination with a physiological model.

• Proceedings of the KSRS Conference
• /
• 2004.10a
• /
• pp.136-139
• /
• 2004

The extraction of 'end-pixels' (i.e. end-members) aims to quantify the abundance of different materials in a single pixel, which becomes popular in the subpixel analysis for hyperspectral dataset. In this paper, we present a new concept called 'End-Pixel of Features (EPF)' to extends the concept of end-pixels for multispectral data and even panchromatic data. The algorithm combines the advantages of previous simplex and clustering methods to search the EPFs in the feature space and reduce the effects of noise. Some experimental results show that, the proposed methodology can be successfully used to hyperspectral data and other remote sensing data.

• Proceedings of the KSRS Conference
• /
• 2008.10a
• /
• pp.243-246
• /
• 2008

The Compact Airborne Imaging Spectrometer System (CAISS) was jointly designed and developed as the hyperspectral imaging system by Korea Aerospace Research Institute (KARI) and ELOP inc., Israel. The primary mission of the CAISS is to acquire and provide full contiguous spectral information with high quality spectral and high spatial resolution for advanced applications in the field of remote sensing. The CAISS consists of six physical units; the camera system, the gyro-stabilized mount, the jig, the GPS/INS, the power inverter and distributor, and the operating system. These subsystems shall be tested and verified in the laboratory before the flight. Especially the camera system of the CAISS shall be calibrated and validated with the calibration equipments such as the integrated sphere and spectral lamps. To improve data quality and availability, it is the most important to understand the mechanism of hyperspectral imaging system and the radiometric and spectral characteristics. This paper presents the major characteristics of camera system on the CAISS and summarizes the results of radiometric and spectral experiment during preliminary system verification.