Pub Date : 2016-08-01DOI: 10.1109/WHISPERS.2016.8071789
Jee-Cheng Wu, Kahn-Bao Wu
The spectrum of each pixel in a hyperspectral image usually comprises multiple material spectra, due to the sensor's spatial resolution and ground material distribution. The purpose of target detection (TD) is to separate specific target pixels from the various background pixels, using a known target signature. In this paper, a novel two-stage target detection process is proposed for improving TD performance. In the first stage, a target detector is applied. In the second stage, the detected result is sorted in ascending order, a portion of the ascending data is selected, and the target detector is reapplied using the selected subset data. In this study, three real hyperspectral data-cubes with ground truth and two well-known target detectors are used to evaluate and compare the performance of this method. The experimental results show that the proposed two-stage TD process improves the detection quality, with a reduced number of false alarms.
{"title":"Two-stage process for improving the performance of hyperspectral target detection","authors":"Jee-Cheng Wu, Kahn-Bao Wu","doi":"10.1109/WHISPERS.2016.8071789","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071789","url":null,"abstract":"The spectrum of each pixel in a hyperspectral image usually comprises multiple material spectra, due to the sensor's spatial resolution and ground material distribution. The purpose of target detection (TD) is to separate specific target pixels from the various background pixels, using a known target signature. In this paper, a novel two-stage target detection process is proposed for improving TD performance. In the first stage, a target detector is applied. In the second stage, the detected result is sorted in ascending order, a portion of the ascending data is selected, and the target detector is reapplied using the selected subset data. In this study, three real hyperspectral data-cubes with ground truth and two well-known target detectors are used to evaluate and compare the performance of this method. The experimental results show that the proposed two-stage TD process improves the detection quality, with a reduced number of false alarms.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"64 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121840370","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-08-01DOI: 10.1109/WHISPERS.2016.8071725
I. Leifer, C. Melton, D. Tratt, Jason Frash, Manish X. Gupta, B. Leen, K. Buckland, P. Johnson
Methane, CH4, and ammonia, NH3, directly and indirectly influence the atmospheric radiative balance. Long wave infrared (LWIR) airborne hyperspectral imagery and in situ data of CH4, CO2, and NH3 plumes were collected from the Chino Dairy Complex in the Los Angeles Basin. LWIR data showed significant emissions heterogeneity between dairies with good spatial agreement with in situ measurements. Remote sensing data also showed topographic effects on plumes mapped for at least 19 km. Repeated in situ measurements showed that emissions were persistent on half-year timescales. Inversion of one dairy plume found annual emissions of 4.1×105 kg CH4, 2.2×105kg NH3, and 2.3×107 kg CO2, suggesting 3500, 4000, and 2100 head of cattle, respectively. Far field data showed chemical conversion of Chino NH3 occurs within the confines of the Los Angeles Basin on % day timescale, faster than previously published values.
甲烷(CH4)和氨(NH3)直接或间接影响大气辐射平衡。利用长波红外(LWIR)机载高光谱图像和CH4、CO2和NH3羽流的原位数据收集了洛杉矶盆地Chino乳业综合体的CH4、CO2和NH3羽流。LWIR数据显示奶牛场之间的排放具有显著的异质性，与原位测量结果具有良好的空间一致性。遥感数据还显示了地形对至少19公里范围内的羽流的影响。重复的现场测量表明，排放在半年的时间尺度上持续存在。对一个奶牛羽流的反演发现，年排放量分别为4.1×105 kg CH4、2.2×105kg NH3和2.3×107 kg CO2，表明奶牛的年排放量分别为3500头、4000头和2100头。远场数据显示，中国NH3的化学转化发生在洛杉矶盆地范围内，以%的时间尺度，比以前公布的值快。
{"title":"Comparing imaging spectroscopy and in situ observations of Chino dairy complex emissions","authors":"I. Leifer, C. Melton, D. Tratt, Jason Frash, Manish X. Gupta, B. Leen, K. Buckland, P. Johnson","doi":"10.1109/WHISPERS.2016.8071725","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071725","url":null,"abstract":"Methane, CH4, and ammonia, NH3, directly and indirectly influence the atmospheric radiative balance. Long wave infrared (LWIR) airborne hyperspectral imagery and in situ data of CH<inf>4</inf>, CO<inf>2</inf>, and NH<inf>3</inf> plumes were collected from the Chino Dairy Complex in the Los Angeles Basin. LWIR data showed significant emissions heterogeneity between dairies with good spatial agreement with in situ measurements. Remote sensing data also showed topographic effects on plumes mapped for at least 19 km. Repeated in situ measurements showed that emissions were persistent on half-year timescales. Inversion of one dairy plume found annual emissions of 4.1×10<sup>5</sup> kg CH<inf>4</inf>, 2.2×10<sup>5</sup>kg NH<inf>3</inf>, and 2.3×10<sup>7</sup> kg CO<inf>2</inf>, suggesting 3500, 4000, and 2100 head of cattle, respectively. Far field data showed chemical conversion of Chino NH<inf>3</inf> occurs within the confines of the Los Angeles Basin on % day timescale, faster than previously published values.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116747474","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-08-01DOI: 10.1109/WHISPERS.2016.8071757
Charis Lanaras, E. Baltsavias, K. Schindler
To enhance the spatial resolution of hyperspectral data, additional multispectral images of higher resolution can be used. However, to combine the two data sources information about the sensors is needed. In this paper we derive a model to estimate the relative spatial and spectral response of the two sensors. The proposed formulation includes non-negativity, recovers remaining registration (shift) errors, and uses prior information to adjust to the shape of the spectral response with either l1 or l2 norm regularization. The framework is tested both with real data and with simulated data where the ground truth is known.
{"title":"Estimation of relative sensor characteristics for hyperspectral super-resolution","authors":"Charis Lanaras, E. Baltsavias, K. Schindler","doi":"10.1109/WHISPERS.2016.8071757","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071757","url":null,"abstract":"To enhance the spatial resolution of hyperspectral data, additional multispectral images of higher resolution can be used. However, to combine the two data sources information about the sensors is needed. In this paper we derive a model to estimate the relative spatial and spectral response of the two sensors. The proposed formulation includes non-negativity, recovers remaining registration (shift) errors, and uses prior information to adjust to the shape of the spectral response with either l1 or l2 norm regularization. The framework is tested both with real data and with simulated data where the ground truth is known.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129800377","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-08-01DOI: 10.1109/WHISPERS.2016.8071740
Snigdha Tariyal, H. Aggarwal, A. Majumdar
In this work we propose a new deep learning tool — deep dictionary learning. We give an alternate neural network type interpretation to dictionary learning. Based on this, we build a deep architecture by cascading one dictionary after the other. The learning proceeds in a greedy fashion, therefore for each level we only need to learn a single layer of dictionary — time tested tools are there to solve this problem. We compare our approach to the deep belief network (DBN) and stacked autoencoder (SAE) based techniques for hyperspectral image classification. We find that in the practical scenario, when the training data is limited, our method outperforms the more established tools like SAE and DBN.
{"title":"Greedy deep dictionary learning for hyperspectral image classification","authors":"Snigdha Tariyal, H. Aggarwal, A. Majumdar","doi":"10.1109/WHISPERS.2016.8071740","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071740","url":null,"abstract":"In this work we propose a new deep learning tool — deep dictionary learning. We give an alternate neural network type interpretation to dictionary learning. Based on this, we build a deep architecture by cascading one dictionary after the other. The learning proceeds in a greedy fashion, therefore for each level we only need to learn a single layer of dictionary — time tested tools are there to solve this problem. We compare our approach to the deep belief network (DBN) and stacked autoencoder (SAE) based techniques for hyperspectral image classification. We find that in the practical scenario, when the training data is limited, our method outperforms the more established tools like SAE and DBN.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126807176","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-08-01DOI: 10.1109/WHISPERS.2016.8071670
K. Uto, Haruyuki Seki, G. Saito, Y. Kosugi, T. Komatsu
The development of monitoring and conservation technology of coastal regions is a key subject to protect the Earth's ecosystem. In the project of “Development of three dimensional mapping system of marine macrophyte beds using hyper- and multispectral remote sensing from air and seasurface” that is supported by JST CREST [1], we develop two mapping systems, i.e., (1) acoustic sensors for measuring water, bottom sediments and the depth of water and (2) hyperspectral imagers for detecting marine macrophytes. In this paper, we investigate the characteristics of remotely sensed hyperspectral images of the north coast of the Izu Oshima, Japan. The hyperspectral images were measured under different illumination condition, i.e., under cloudy and sunny skies, baed on a whiskbroom hyperspectral imager [2].
{"title":"Measurement of a coastal area by a hyperspectral imager using an optical fiber bundle, a swing mirror and compact spectrometers","authors":"K. Uto, Haruyuki Seki, G. Saito, Y. Kosugi, T. Komatsu","doi":"10.1109/WHISPERS.2016.8071670","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071670","url":null,"abstract":"The development of monitoring and conservation technology of coastal regions is a key subject to protect the Earth's ecosystem. In the project of “Development of three dimensional mapping system of marine macrophyte beds using hyper- and multispectral remote sensing from air and seasurface” that is supported by JST CREST [1], we develop two mapping systems, i.e., (1) acoustic sensors for measuring water, bottom sediments and the depth of water and (2) hyperspectral imagers for detecting marine macrophytes. In this paper, we investigate the characteristics of remotely sensed hyperspectral images of the north coast of the Izu Oshima, Japan. The hyperspectral images were measured under different illumination condition, i.e., under cloudy and sunny skies, baed on a whiskbroom hyperspectral imager [2].","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127893460","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-08-01DOI: 10.1109/WHISPERS.2016.8071673
P. Adams, D. Lynch, K. Buckland, P. Johnson, D. Tratt
Several ammonia emitting fumarole fields have recently been exposed along the southeastern shoreline of the Salton Sea in Imperial County, California. A complex assemblage of sulfate minerals, many containing ammonium ion, are associated with the fumaroles. The distribution of these sulfates was mapped by remote sensing with the Mako LWIR hyperspectral sensor. The most common minerals tended to form somewhat concentric discontinuous rings. Outwardly from the central fumarole vent, they were: mascagnite/boussingaultite, gypsum, nitratine and bloedite, respectively. Ground truth surveys coupled with laboratory analyses were generally in good agreement with the remote sensing results.
{"title":"Hyperspectral LWIR mapping of fumarole sulfates, salton sea, imperial county, California","authors":"P. Adams, D. Lynch, K. Buckland, P. Johnson, D. Tratt","doi":"10.1109/WHISPERS.2016.8071673","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071673","url":null,"abstract":"Several ammonia emitting fumarole fields have recently been exposed along the southeastern shoreline of the Salton Sea in Imperial County, California. A complex assemblage of sulfate minerals, many containing ammonium ion, are associated with the fumaroles. The distribution of these sulfates was mapped by remote sensing with the Mako LWIR hyperspectral sensor. The most common minerals tended to form somewhat concentric discontinuous rings. Outwardly from the central fumarole vent, they were: mascagnite/boussingaultite, gypsum, nitratine and bloedite, respectively. Ground truth surveys coupled with laboratory analyses were generally in good agreement with the remote sensing results.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"67 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126305948","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Like most spectral remote sensing data, ASTER images reflect different spectral information of surface objects, and ground-based magnetic data reflect magnetic information from the surface and from rocks at depth. In this study, a magnetic image was first generated and then combined with ASTER spectral bands to provide multispectral data. The minimum distance and maximum likelihood methods were used to classify lithology mapping in the Zhalute area. Classification results show that, individually, spectral and magnetic data have advantages for some aspects of lithological mapping but the integrated spectrum-magnetic data improved overall accuracy. This study shows that the integrated use of ASTER data and magnetic data has useful applications for the field of lithological mapping.
{"title":"Lithological mapping using ASTER and magnetic data: A case study from Zhalute area, China","authors":"Jiang Chen, Qun Zhu, Weijun Zhao, Zhongren Sun, Chunpeng Zhang, Zhaoxia Mao, Qian Zhao","doi":"10.1109/WHISPERS.2016.8071803","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071803","url":null,"abstract":"Like most spectral remote sensing data, ASTER images reflect different spectral information of surface objects, and ground-based magnetic data reflect magnetic information from the surface and from rocks at depth. In this study, a magnetic image was first generated and then combined with ASTER spectral bands to provide multispectral data. The minimum distance and maximum likelihood methods were used to classify lithology mapping in the Zhalute area. Classification results show that, individually, spectral and magnetic data have advantages for some aspects of lithological mapping but the integrated spectrum-magnetic data improved overall accuracy. This study shows that the integrated use of ASTER data and magnetic data has useful applications for the field of lithological mapping.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130107583","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-08-01DOI: 10.1109/WHISPERS.2016.8071662
Sathishkumar Samiappan, Lalitha Dabbiru, R. Moorhead
Hyperspectral imagery provides detailed information about land-cover materials over a wide spectral range. Land-cover classification using hyperspectral data has been an active topic of research. Elevation data from light detection and ranging (LiDAR) can aid the classification process in discriminating complex classes. Fusion of hyperspectral and LiDAR data has been investigated in the past where the goal was to extract features from both sources and combine them to improve the accuracy of land-cover classification. In this paper, we present a new fusion approach based on random feature selection (RFS) and morphological attribute profiles (AP). Our experimental study, conducted on a hyperspectral image and digital surface model (DSM) derived from first return LiDAR data collected over the Samford ecological research facility, Queensland, Australia indicate that the proposed approach yields excellent classification results.
{"title":"Fusion of hyperspectral and LiDAR data using random feature selection and morphological attribute profiles","authors":"Sathishkumar Samiappan, Lalitha Dabbiru, R. Moorhead","doi":"10.1109/WHISPERS.2016.8071662","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071662","url":null,"abstract":"Hyperspectral imagery provides detailed information about land-cover materials over a wide spectral range. Land-cover classification using hyperspectral data has been an active topic of research. Elevation data from light detection and ranging (LiDAR) can aid the classification process in discriminating complex classes. Fusion of hyperspectral and LiDAR data has been investigated in the past where the goal was to extract features from both sources and combine them to improve the accuracy of land-cover classification. In this paper, we present a new fusion approach based on random feature selection (RFS) and morphological attribute profiles (AP). Our experimental study, conducted on a hyperspectral image and digital surface model (DSM) derived from first return LiDAR data collected over the Samford ecological research facility, Queensland, Australia indicate that the proposed approach yields excellent classification results.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130446695","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-08-01DOI: 10.1109/WHISPERS.2016.8071738
K. Westberg, Jeffrey E. Matic
The Aerospace Corporation's portable chemical release equipment has been used for some time to discharge gases and atomized liquids into the atmosphere at accurately measured flow rates, thereby producing chemical plumes whose column densities can be determined by remote infrared imaging spectrometers. Column densities can be converted into mass flow rates with a knowledge of the wind speed, the air temperature, and the ground and/or the sky radiometric temperature, which are also measured at the chemical release site, simultaneously with the release. Chemical releases have been, and continue to be, used to determine the smallest chemical plume that can be detected by an imaging spectrometer under varying conditions and to determine the accuracy to which it can infer chemical flow rates, air temperature, and wind speed. This paper describes the equipment and procedures used to release chemicals into the atmosphere and make the required meteorological and radiometric temperature measurements. The accuracy of each measurement is given.
{"title":"Generating chemical plumes for imaging spectrometers: Equipment and procedures","authors":"K. Westberg, Jeffrey E. Matic","doi":"10.1109/WHISPERS.2016.8071738","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071738","url":null,"abstract":"The Aerospace Corporation's portable chemical release equipment has been used for some time to discharge gases and atomized liquids into the atmosphere at accurately measured flow rates, thereby producing chemical plumes whose column densities can be determined by remote infrared imaging spectrometers. Column densities can be converted into mass flow rates with a knowledge of the wind speed, the air temperature, and the ground and/or the sky radiometric temperature, which are also measured at the chemical release site, simultaneously with the release. Chemical releases have been, and continue to be, used to determine the smallest chemical plume that can be detected by an imaging spectrometer under varying conditions and to determine the accuracy to which it can infer chemical flow rates, air temperature, and wind speed. This paper describes the equipment and procedures used to release chemicals into the atmosphere and make the required meteorological and radiometric temperature measurements. The accuracy of each measurement is given.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130992630","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Hyperspectral remote sensing image (HSI) consists of hundreds of bands that contain rich space, radiation and spectral information. The high-dimensional data can also lead to the course of dimensionality problem making it difficult to be used effectively. In this paper, we proposed a manifold learning algorithm to reduce the dimensionality for HSI data. For high dimensional datasets with continuous variables, it is often the case that the data points are arranged along with low dimensional structures, named manifolds, in the high dimensional space. Manifold learning aims to identifying those special low dimensional structures for subsequent usage such as classification or regression. However, many manifold learning algorithms perform an eigenvector analysis on a data similarity matrix whose size is N∗N, where N is the number of data points. The memory complexity of the analysis is at least O(N2) that is not feasible for a regular computer to compute or storage for very large datasets. To solve this problem, we used statistical sampling methods to sample a subset of data points as landmarks. A skeleton of the manifold was then identified based on the landmarks. The remaining data points were then inserted into the skeleton by Locally Linear Embedding (LLE). We tested our algorithm on AVIRIS Salinas-A data set. The experimental results showed that the HSI dataset could be reduced to a lower-dimensional space for land use classification with good performance, and the main structure was preserved well.
{"title":"A novel manifold learning for dimensionality reduction and classification with hyperspectral image","authors":"Zezhong Zheng, Pengxu Chen, Mingcang Zhu, Zhiqin Huang, Yufeng Lu, Yicong Feng, Jiang Li","doi":"10.1109/WHISPERS.2016.8071697","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071697","url":null,"abstract":"Hyperspectral remote sensing image (HSI) consists of hundreds of bands that contain rich space, radiation and spectral information. The high-dimensional data can also lead to the course of dimensionality problem making it difficult to be used effectively. In this paper, we proposed a manifold learning algorithm to reduce the dimensionality for HSI data. For high dimensional datasets with continuous variables, it is often the case that the data points are arranged along with low dimensional structures, named manifolds, in the high dimensional space. Manifold learning aims to identifying those special low dimensional structures for subsequent usage such as classification or regression. However, many manifold learning algorithms perform an eigenvector analysis on a data similarity matrix whose size is N∗N, where N is the number of data points. The memory complexity of the analysis is at least O(N2) that is not feasible for a regular computer to compute or storage for very large datasets. To solve this problem, we used statistical sampling methods to sample a subset of data points as landmarks. A skeleton of the manifold was then identified based on the landmarks. The remaining data points were then inserted into the skeleton by Locally Linear Embedding (LLE). We tested our algorithm on AVIRIS Salinas-A data set. The experimental results showed that the HSI dataset could be reduced to a lower-dimensional space for land use classification with good performance, and the main structure was preserved well.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128813938","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：