Earticle

본 연구는 보물 백담사 저고리에 사용된 직물 문양 분석에 대한 단파적외선 (SWIR) 초분광 영상 기술에 관한 것이다. 연구 대상이 되는 보물 백담사 저고리는 1748년이라는 정확한 연대를 알 수 있어, 복식 연구에 있어 매우 중요한 유물이다. 그러나 1993년 보물로 지정되었음에도 불구하 고, 황색의 문양 직물을 제외하면, 자주색 삼회장 부분의 문양 정보를 확보하지 못했다. 사실 황색에 비해 자주색은 카메라가 인식하기 어려운 색이므로, 고화질의 디지털 촬영이나 적외선 촬영만으로는 어려움이 있다. 이에 최근에 문화재 현장에서 각광받고 있는 단파적외선 초분광 영상 촬영 기술을 활용하여, 보물 백담사 저고리의 직물 문양을 식별하는 데에 적용하였다. 그 결과, 만자화문단과 용봉 문의 문양을 확인할 수 있었다. 우리나라의 경우, 이러한 직물 유형은 17세기 말에서 18세기 초중반 에 확인되는 것이다. 금번의 조사는 수년간 확인할 수 없었던 백담사 저고리의 직물 문양의 상세 이미지를 확보했을 뿐만 아니라 직물문화재 연구에 초분광 영상 기술의 새로운 활용법을 제시하였다 는데 의의가 있다.

This study applied short-wave infrared (SWIR) hyperspectral imaging technology to analyze the textile pattern used in the Jeogori of Baekdamsa Temple. First introduced in 1748, the Jeogori is a v ery important relic in t he s tudy o f Korean t raditional costume. T he g arment w as designated as a national treasure in 1993, and the pattern in its yellow silk has been well established. However, obtaining the pattern information of its reddish purple decoration has proven to be quite challenging. In contrast to yellow, reddish purple is a color that is difficult for a camera to recognize; thus, analyzing patterns in this color using only high-definition digital or infrared photography is difficult. In this work, SWIR hyperspectral imaging technology, which has recently been used in investigations of cultural heritage sites, was applied to identify the textile pattern on the Jeogori of Baekdamsa Temple. The technology successfully identified the patterns of a Wan-character, dragon, and phoenix. This type of textile pattern was identified in Korea from the end of the 17th century to the early-to-mid-18th century. This study is meaningful because it not only provides a detailed image of the textile pattern on the Jeogori of Baekdamsa Temple, which has not been previously achieved, but also proposes a new strategy based on hyperspectral imaging technology for studying the cultural heritage of textiles.

하천 주변 친수구역 조성, 4대강 사업 등과 같은 하천정비 사업으로 인해 하천의 흐름특성은 계속적으로 변동하고 있으며, 각종 오염물질 유입으로 인한 수질사고의 위험이 높아지고 있다. 수질사고 발생시 하천의 흐름특성을 고려해 오염물질의 농도 및 도달시간을 예측해 신속한 방제작업으로 하류로의 영향을 최소화해야한다. 이러한 오염물질의 거동을 추적하기 위해서는 하천의 구간별 확산계수, 분산계수 산정이 필요하며 그중 분산계수는 용존성 오염물질의 확산범위 해석에 사용된다. 오염물질의 거동을 추적하기 위한 기존 실험적 연구사례들은 많은 인력과 비용이 소요되고, 한정적인 장비의 운용으로 공간적으로 높은 해상도의 자료 취득이 어려웠다. 최근에는 RGB드론을 이용한 오염물질의 추적연구가 수행되었지만, RGB영상 역시 분광정보를 한정적으로 수집한다는 한계가 있다. 본 연구에서는 기존 연구들의 한계점들을 보완하기 위해 드론을 활용한 원격탐사 플랫폼에 초분광센서를 탑재하여 기존 접촉식 측정보다 시간적, 공간적으로 고해상도의 자료를 수집하였다. 수집된 시공간(Spatio-temporal) 초분광영상을 활용해 추적자의 농도를 산정하고, 횡분산계수를 도출하였다. 향후 연구를 통해 드론 플랫폼의 한계를 극복하고, 분산계수 산정 기술을 고도화하면 수계로 유출되는 각종 오염물질의 감지 및 다양한 수질항목 및 하천인자의 변화량 감지가 가능할 것으로 기대된다.

Due to river maintenance projects such as the creation of hydrophilic areas around rivers and the Four Rivers Project, the flow characteristics of rivers are continuously changing, and the risk of water quality accidents due to the inflow of various pollutants is increasing. In the event of a water quality accident, it is necessary to minimize the effect on the downstream side by predicting the concentration and arrival time of pollutants in consideration of the flow characteristics of the river. In order to track the behavior of these pollutants, it is necessary to calculate the diffusion coefficient and dispersion coefficient for each section of the river. Among them, the dispersion coefficient is used to analyze the diffusion range of soluble pollutants. Existing experimental research cases for tracking the behavior of pollutants require a lot of manpower and cost, and it is difficult to obtain spatially high-resolution data due to limited equipment operation. Recently, research on tracking contaminants using RGB drones has been conducted, but RGB images also have a limitation in that spectral information is limitedly collected. In this study, to supplement the limitations of existing studies, a hyperspectral sensor was mounted on a remote sensing platform using a drone to collect temporally and spatially higher-resolution data than conventional contact measurement. Using the collected spatio-temporal hyperspectral images, the tracer concentration was calculated and the transverse dispersion coefficient was derived. It is expected that by overcoming the limitations of the drone platform through future research and upgrading the dispersion coefficient calculation technology, it will be possible to detect various pollutants leaking into the water system, and to detect changes in various water quality items and river factors.

The objective of this study was to classify between radish and Chinese cabbage in autumn using hyperspectral images. The hyperspectral images were acquired by Compact Airborne Spectrographic Imager (CASI) with 1m spatial resolution and 48 bands covering the visible and near infrared portions of the solar spectrum from 370 to 1044 nm with a bandwidth of 14 nm. An object-based technique is used for classification of radish and Chinese cabbage. It was found that the optimum parameter values for image segmentation were scale 400, shape 0.1, color 0.9, compactness 0.5 and smoothness 0.5. As a result, the overall accuracy of classification was 90.7 % and the kappa coefficient was 0.71. The hyperspectral images can be used to classify other crops with higher accuracy than radish and Chines cabbage because of their similar characteristic and growth time.

본 연구의 핵심은 하이퍼스펙트럴영상에 정준상관분류기법을 적용할 때, 최적의 분광밴드를 찾아내는 유효밴드 선정 및 추출기법은 무엇인가를 알아내는 것이다. 본 연구에서는 미국의 Purdue University에서 개발된 Multispec$^{(C)}$ 소프트웨어를 사용하여 각각의 분리도 결정기법에 따른 최적의 유효밴드를 선정하였다. 사용된 분리도 결정기법은 Divergence, Transformed Divergence, Bhattacharyya, Mean Bhattacharyya, Covariance Bhattacharyya, Non Covariance Bhattacharyya로서 총 6가지이다. 특징추출을 위해 Erdas Imagine과 ENVI 소프트웨어를 사용하여 PCA 변환과 MNF 변환을 수행하였다. 유효밴드 선정 및 특징추출의 효과에 대한 비교평가를 위해, 정준상관분류기법에 의한 토지피복분류작업을 수행하였다. 1차 선별된 60개 밴드를 사용한 정준상관분류의 정확도는 71.8%이며, 정준상관분류를 사용하여 가장 높은 분류정확도를 얻은 방법은 Noncovariance Bhattacharyya 적용 후 정준상관분류를 수행한 경우로서 전체정확도 79.0% 이다. 결론적으로 정준상관분류에 의한 하이퍼스펙트럴영상 분류에서는 유효밴드선정기법으로 사실상 Noncovariance Bhattacharyya 기법만 유용하였으며, 나머지 유효밴드 선정기법(Divergence 제외)과 특징추출기법은 정준상관분류에서는 오히려 분류정확도가 하락함을 확인하였다.

The core of this study is finding out the efficient band selection or extraction method discovering the optimal spectral bands when applying canonical correlation classifier (CCC) to hyperspectral data. The optimal efficient bands grounded on each separability decision technique are selected using Multispec$^{(C)}$ software developed by Purdue university of USA. Total 6 separability decision techniques are used, which are Divergence, Transformed Divergence, Bhattacharyya, Mean Bhattacharyya, Covariance Bhattacharyya, Noncovariance Bhattacharyya. For feature extraction, PCA transformation and MNF transformation are accomplished by ERDAS Imagine and ENVI software. For the comparison and assessment on the effect of feature selection and feature extraction, land cover classification is performed by CCC. The overall accuracy of CCC using the firstly selected 60 bands is 71.8%, the highest classification accuracy acquired by CCC is 79.0% as the case that executes CCC after appling Noncovariance Bhattacharyya. In conclusion, as a matter of fact, only Noncovariance Bhattacharyya separability decision method was valuable as feature selection algorithm for hyperspectral image classification depended on CCC. The lassification accuracy using other feature selection and extraction algorithms except Divergence rather declined in CCC.

하천법 개정 및 수자원의 조사·계획 및 관리에 관한 법률 제정으로 하상변동조사를 정기적으로 실시하는 것이 의무화되었고, 지자체가 계획적으로 수자원을 관리할 수 있도록 제도가 마련되고 있다. 하상 지형은 직접 측량할 수 없기 때문에 수심 측량을 통해 간접적으로 이루어지고 있으며, 레벨측량이나 음향측심기를 활용한 접촉식 방법으로 이루어지고 있다. 접촉식 수심측량법은 자료수집이 제한적이기 때문에 공간해상도가 낮고 연속적인 측량이 불가능하다는 한계가 있어 최근에는 LiDAR나 초분광영상을 이용한 원격탐사를 이용한 수심측정 기술이 개발되고 있다. 개발된 초분광영상을 이용한 수심측정 기술은 접촉식 조사보다 넓은 지역을 조사할 수 있고, 잦은 빈도로 자료취득이 용이한 드론에 경량 초분광센서를 탑재하여 초분광영상을 취득하고, 최적 밴드비 탐색 알고리즘을 적용해 수심분포 산정이 가능하다. 기존의 초분광 원격탐사 기법은 드론의 경로비행으로 획득한 초분광영상을 면단위의 영상으로 정합한 후 특정 물리량에 대한 분석이 수행되었으며, 수심측정의 경우 모래하천을 대상으로 한 연구가 주를 이루었으며, 하상재료에 대한 평가는 이루어지지 않았었다. 본 연구에서는 기존의 초분광영상을 활용한 수심산정 기법을 식생이 있는 하천에 적용하고, 동일지역에서 식생을 제거한 후의 2가지 케이스에 대해서 시공간 초분광영상과 단면초분광영상에 모두 적용하였다. 연구결과, 식생이 없는 경우의 수심산정이 더 높은 정확도를 보였으며, 식생이 있는 경우에는 식생의 높이를 바닥으로 인식한 수심이 산정되었다. 또한, 기존의 단면초분광영상을 이용한 수심산정뿐만 아니라 시공간 초분광영상에서도 수심산정의 높은 정확도를 보여 시공간 초분광영상을 활용한 하상변동(수심변동) 추적의 가능성을 확인하였다.

Due to the revision of the River Act and the enactment of the Act on the Investigation, Planning, and Management of Water Resources, a regular bed change survey has become mandatory and a system is being prepared such that local governments can manage water resources in a planned manner. Since the topography of a bed cannot be measured directly, it is indirectly measured via contact-type depth measurements such as level survey or using an echo sounder, which features a low spatial resolution and does not allow continuous surveying owing to constraints in data acquisition. Therefore, a depth measurement method using remote sensing-LiDAR or hyperspectral imaging-has recently been developed, which allows a wider area survey than the contact-type method as it acquires hyperspectral images from a lightweight hyperspectral sensor mounted on a frequently operating drone and by applying the optimal bandwidth ratio search algorithm to estimate the depth. In the existing hyperspectral remote sensing technique, specific physical quantities are analyzed after matching the hyperspectral image acquired by the drone's path to the image of a surface unit. Previous studies focus primarily on the application of this technology to measure the bathymetry of sandy rivers, whereas bed materials are rarely evaluated. In this study, the existing hyperspectral image-based water depth estimation technique is applied to rivers with vegetation, whereas spatio-temporal hyperspectral imaging and cross-sectional hyperspectral imaging are performed for two cases in the same area before and after vegetation is removed. The result shows that the water depth estimation in the absence of vegetation is more accurate, and in the presence of vegetation, the water depth is estimated by recognizing the height of vegetation as the bottom. In addition, highly accurate water depth estimation is achieved not only in conventional cross-sectional hyperspectral imaging, but also in spatio-temporal hyperspectral imaging. As such, the possibility of monitoring bed fluctuations (water depth fluctuation) using spatio-temporal hyperspectral imaging is confirmed.

분뇨 액비는 가축 사육시 발생되는 분뇨를 효율적으로 재사용하기 위한 대안이 되고 있으나, 특정 지점에 대량 살포 되거나 무분 별하게 사용할 경우 토양과 지하수 오염을 발생 시킬 수 있다. 본 논문에서는 액비 핫스팟을 탐지하고자 비파괴 분석이 가능하고 정밀한 스펙트럼 분석이 사능한 초분광계를 이용하여 연구를 진행했다. 제주도 현지에서 오염되지 않은 토양시료와 액비가 대량 살포된 토양시료를 채집하여 5단계(0%, 25%, 50%, 75%, 100%)의 액비 농도별 시료를 배합했다. 먼저 액비 농도별 스펙트럼을 분석하고자 점단위로 초분광 데이터를 획득할 수 있는 Fieldspec4 장비(400 ~ 2500 nm)를 이용하여 액비 농도별 스펙트럼 변화추 세를 확인했다. 액비에 함유된 중금속 성분 중 아연에 대한 농도를 추정 하기위해 면단위 초분광 영상를 획득할 수 있는 FX17e 초분 광 카메라(900 ~ 1700 nm)를 이용했고, 획득한 초분광 영상에 PLSR(Partial Least Squares Regression) 회귀분석 방법을 적용하였 다. PLSR 분석의 종속변수는 채집한 액비시료를 보건환경연구원에 의뢰하여 얻은 실제 아연 농도값, 독립변수는 액비 농도별 초분 광 영상의 반사율 값이다. 두 변수의 상관관계를 모델링하여 회귀식을 구한 후 예측모델을 개발했고, RMSE(Root Mean Square Error)는 0.91로 높은 상관성을 보였다. 모델을 통해 추정한 아연농도의 맵 분포도를 생성했을 때 핫스팟 추정이 가능하다고 판단하 였고, 추후 본 연구 결과를 토대로 토양 액비 오염탐지를 위한 현장실증에 대한 연구가 진행될 예정이다.

이 연구에서는 하이퍼스펙트럴 이미지를 활용하여 석조문화재의 훼손상태 모니터링에 대한 활용성을 검토하였다. 이를 위해 하이퍼스펙트럴 데이터의 보정방법, 영상분류 및 정규화 식생지수 산출방법을 석조문화재에 적용하였다. 이 결과 각 물질의 분광정보를 기반으로 한 객관적인 훼손지도 작성, 정밀도 높은 훼손율의 산출 및 식생들의 활력도모델작성 등 다양한 분석이 가능하였다. 따라서 하이퍼스펙트럴 이미지 분석을 활용하여 석조문화재를 모니터링 한다면 효율적으로 훼손상태 변화를 파악할 수 있을 것이다.

This study was considered utilization of hyperspectral image analysis for monitoring. Accordingly we applied to stone cultural properties to data correction methods, image classification techniques, NDVI computation techniques using hyperspectral image. As the results, hyperspectral image analysis was possible making detailed deterioration map, accurate calculation of deterioration rate, mapping of normalized difference vegetation index on the basis of reflectance of each materials. Therefore, hyperspectral image analysis will be used for effective monitoring techniques of stone cultural heritages.

초분광 이미지 분석은 석조문화재 손상지도 제작 시 화학적⋅생물학적 오염의 정량적인 면적 산출이 어려운 단점을 보완하기 위해 제안되었다. 이 연구에서는 다양한 표면오염물이 나타나는 경주 굴불사지 석조사면불상을 대상으로 초분광 이미지 분석을 수행하였다. 이때 화학적⋅생물학적 오염 은 색상과 형태에 따라 10가지 범주로 구분하였고 범주의 참조 이미지 제작법을 제시하였다. 또한 오염물을 분류하기 위해 SAM 알고리듬을 사용하고 관심영역을 사용한 분류방법(Method A)과 영상 에서 추출한 분광정보를 라이브러리로 구축하여 분류하는 방법(Method B)을 적용하였다. 분류된 이 미지를 참조 이미지와 비교한 결과, 정확도는 분류 방법에 따라 52.07%와 63.61%, Kappa 지수는 0.43과 0.55로 산출되었고, 분류시 오분류 화소는 동일한 계열의 오염에 분포하는 경향을 보인다.

Considering the difficulties associated with the creation of deterioration maps for stone cultural heritages, quantitative determination of chemical and biological contaminants in them is still challenging. Hyperspectral image analysis has been proposed to overcome this drawback. In this study, hyperspectral imaging was performed on Stone Buddhas Temple in Four Directions at Gulbulsa Temple Site(Treasure 121), and several surface contaminants were observed. Based on the color and shape, these chemical and biological contaminants were classified into ten categories. Additionally, a method for establishing each class as a reference image was suggested. Simultaneously, with the help of Spectral Angle Mapper algorithm, two classification methods were used to classify the surface contaminants. Method A focused on the region of interest, while method B involved the application of the spectral library prepared from the image. Comparison of the classified images with the reference image revealed that the accuracies and kappa coefficients of methods A and B were 52.07% and 63.61%, and 0.43 and 0.55, respectively. Additionally, misclassified pixels were distributed in the same contamination series.

Hyperspectral image mostly have very large amounts of data which makes the computational cost and subsequent classification task a difficult issue. Firstly, to solve the problem of computational complexity, spectral clustering algorithm is imported to select efficient bands for subsequent classification task. Secondly, due to lack of labeled training sample points, this paper proposes a new algorithm that combines support vector machines and Bayesian classifier to create a discriminative/generative hyperspectral image classification method using the selected features. Experimental results on real hyperspectral image show that the proposed method has better performance than the other state-of-the-art methods.

The abundant information available in hyperspectral image has provided important opportunities for land-cover classification and recognition. However, “Curse of dimensionality” and small training sample set are two difficulties which hinder the improvement of computational efficiency and classification precision. In this paper, we present a co-training based method on hyperspectral image classification. Firstly, two views of samples are generated through two kinds of dimensionality reduction methods. After that, the co-training process is viewed as combinative label propagation over two independent views. Experimental results on real hyperspectral image show that the proposed method has better performance than the other state-of-the-art methods.

The limited resolution of image sensors and the complex diversity of nature, cause mixed pixel problems in hyperspectral technology. Such problems are common, and increase the complexity of hyperspectral image processing. Hyperspectral unmixing is crucial for hyperspectral image classification and recognition. In unmixing, the image signatures are represented as a linear combination of the basic materials. Unmixing is the process of decomposing a mixed pixel into constituent materials, and calculating the corresponding fractional abundance. If pure materials (end members) are present in an image, unmixing can be divided into two steps, namely, end member extraction and abundance decomposition. On the other hand, if there is no pure material, researchers have devised and investigated unsupervised and semi-supervised spectral unmixing technology. This article presents an overview of the state-of-the-art methods of hyperspectral unmixing and their extensions.

According to the characteristics of hyperspectral images, a novel compression and reconstruction algorithm for hyperspectal images based on compressed sensing is proposed. The random measurements of each image and the linear prediction coefficients are made at the encoder, and then transmitted sequentially to the decoder. At the decoder, in terms of apparent correlations between the adjacent spectral bands, a de-correlation algorithm based on block linear prediction model is used in reconstruction process. The inter-band redundancies are removed from the measurements of current image, thus the de-correlation image data is sparser, which can be reconstructed easily. Experimental results show that the proposed algorithm achieves improved reconstruction performance and efficiently reduces the cost of computation at the encoder, which is more suitable for hardware implementation.

Hyperspectral image classification has received an increasing amount of interest in recent years. However, when representing pixels as vectors, the dimensionality of feature space is high, which causes “curse of dimensionality” problem. In this paper, in order to alleviate the impact of above problem, a manifold sparse coding method is proposed. Firstly, matrix decomposition technique is used to find a concept set and calculates relative data projection in the concept set. Secondly, manifold learning regularization is imported into objective function to capture the intrinsic geometric structure in the data. Finally, LASSO regularization is used to obtain sparse representation of data projection. Experimental results on real hyperspectral image show that the proposed method has better performance than the other state-of-the-art methods.

With the development of spectral imaging technology, it makes hyperspectral imagery widely used. According to the features of multiple bands and the strong mutual correlation among these bands, this paper presents a band selection method for hyperspectral imagery classification based on improved PSO (Particle Swarm Optimization). First of all, we use information divergence to describe the correlation of the bands, then build the information divergence matrix to make the classification of subspaces. Secondly, we construct the fitness function of the algorithm with the band information and categories of the Bhattacharyya distance (B distance) to improve the inertia weight updating method in PSO. Finally, based on the AVIRIS hyperspectral imagery and compared with existing method to conduct experiments, the average classification accuracy of the proposed method is 81.36%, which is distinctly improved 0.91% compared with the existed method. Meanwhile, the proposed method has a significantly faster convergence speed during the process of the band selection. Therefore, the experimental results verify the effectiveness of the proposed method in this paper.

The study on classification methods of hyperspectral image is a focal growing area in remote sensing applications because the wide spectral range, providing a very high spectral resolution, allows the detection and classification surfaces and chemical elements of the observed image. Semi-supervised learning method which takes a large number of unlabeled samples and minority labeled samples, improving classification and predicting the accuracy effectively have been a new research direction. In this paper we proposed a new semi-supervised classification method of hyperspectral image based on combining Renyi entropy and multinomial logistic regression algorithm. The multinomial logistic regression was performed to describe a direct relationship between the selected sample as and their category. A lot of unlabeled samples are constantly added to the sample data using Renyi entropy algorithm. The test analysis of image classification in test area showed the advantages of classification method based on combining Renyi entropy and multinomial logistic regression algorithm for hyperspectral remote sensing image.

Obtaining thematic maps using image classification techniques from hyperspectral datasets is a very difficult image processing task. In hyperspectral image analysis dimensionality reduction is one of the challenging pre-processing tasks, which is achieved using feature extraction techniques. The beauty of these techniques is that they drastically reduces the dimensionality of image and at the same time preserves the majority of the essential information. In this paper few most frequently used dimensionality reduction methods are being investigated, which helps to get accurateness. This research work presents a relative performance investigation of few mostly frequently used feature extraction techniques like Decision Boundary Feature Extraction (DBFE), Non-Parametric Weighted Feature Extraction (NWFE), Discriminative analysis feature extraction (DAFE) and Principal Component Analysis (PCA). The classification is carried out using two most widely used classification techniques including Gaussian Maximum Likelihood (GML) and neural network (NNs). The results obtained after performing experiments indicates that Decision Boundary Feature Extraction (DBFE) technique has provided the best accuracy among various investigated feature extraction techniques. The application areas of this research include areas like identification of exact location in battle field, drought affected areas, flooded areas and weather forecasting etc.

This paper presents a new technique for marker selection called marker selection using skeletonization. Markers are the most reliable pixels that represent a particular class. Marker selection using skeletonization is further analysed to do classification of hyperspectral image with very low training samples, as low as one pixel per class. Both spatial and spectral information are used to improve the final classification accuracy. An Extended Morphological Profile with duality is used to extract spatial information. Furthermore, it is shown that by using the spatial and spectral information with non- parametric supervised feature extraction methods, better classification accuracy can be achieved even when very low training samples are available. The classification maps will be shown and discussed for very low training sample analysis using marker selection by skeletonization technique.

In this paper we have proposed a new marker selection technique using Support Vector Machine over-fitting. Markers are the most reliable pixels in a class. We used our proposed technique to do classification of hyperspectral image with very low training samples, as low as one pixel per class. We have used both spectral and spatial information to improve the classification results. The spatial information is extracted using Extended Morphological Profiles with duality. Nonparametric supervised feature extraction methods are used to eliminate the redundant and irrelevant information in both spatial and spectral domains. In the end we have done experimentation to verify our proposed approach. The experimentation results show that when non-parametric weighted feature extraction method is used we get better classification results. The classification maps shows that even with just one training sample per class we still can get a reliably reasonable classification map.

For hyperspectral remote sensing image denoising, this paper proposed image denoising based on non-local low-rank dictionary learning. The basic idea of algorithm is to use strong relativity of all wave bands of hyperspectral remote sensing image with local self-similarity and local sparsity of image to improve the denoising performance. First of all, combined with the strong relativity, non-local self-similarity and local sparsity, non-local low-rank dictionary learning is established. Then iterative method is used to solve the model to get redundant dictionary and sparsity to represent coefficient. Finally, redundant dictionary and sparsity is used to express restored image of coefficient. Compared with the existing advanced algorithm, by making full use of strong relativity each band of hyperspectral image, it makes the algorithm obtain the information on details to well keep the hyperspectral remote sensing image, to improve the visual effect. Experimental results verify the effectiveness of the algorithm in this paper.

초분광 이미지는 일반 이미지와 달리 전자기 스펙트럼을 파장에 따라 수많은 밴드로 나누어 촬영된 것으로 고용 량 고해상도 이미지이다. 일반 이미지보다 정보량이 많아 물체나 물질 탐사에 활용된다. 처리할 초분광 이미지의 정보량 을 줄이기 위해 밴드 선택(band selection)기법[3]을 활용한다. 기존 밴드 선택기법들은 통계를 바탕으로 하는 휴리스틱 한 기법으로, 시간이 오래 걸리며, 일반성과 보편성이 떨어지는 경우가 많다. 이를 보완하기 위해 본 논문에서는 양자화 개념(Quantization)를 활용하여, 이산 범위(Discrete Range)를 통해 범위별로 대표적인 밴드를 뽑아 밴드 선택에 사용 한다. 실험 결과를 통해 제안 기법이 기존 밴드 선택 방식보다 수행 시간이 매우 빠르며 밴드 수를 1/10~1/7로 줄였음 에도 원본과 성능 정확도가 유사함을 보였다.

Unlike or common images, Hyperspectral images were taken by continuous electromagnetic spectral into numerous bands according to wavelengths and are high-capacity high-resolution images. It has more information than ordinary images, so it is used to explore objects and materials. To reduce the amount of information in hyper-spectral images to be processed, band selection is utilized. Existing band selection techniques are heuristic techniques based on statistics, which take a long time and often lack generality and universality. To compensate for this, this paper utilizes quantization concept to draw representative bands through Discrete Range, we use them for band selection algorithm. Experimental results showed that the proposed technique performed much faster than conventional band selection methods, and that the performance accuracy was similar to that of the original even though the number of bands was reduced by one-seventh to one-tenth.