Pub Date : 2016-08-01DOI: 10.1109/WHISPERS.2016.8071802
Yuan Zhou, Anand Rangarajan, P. Gader
Endmember variability complicates the problem of spectral unmixing. This variability is typically represented by probability distributions or spectral libraries. The present work describes a new distributional representation based on Gaussian Mixture Models (GMMs). The most common form in this setting is the Normal Compositional Model (NCM), wherein the endmembers for each pixel are modeled as samples drawn from unimodal Gaussians. In reality, however, the distribution of spectra from a material may be multi-modal. We first show that a linear combination of GMM random variables is also a GMM. This allows us to probabilistically formulate hyperspectral pixel likelihoods as combinations of independent endmember random variables. Then, after adding a reasonable smoothness and sparsity prior on the abundances, we obtain a standard Bayesian maximum a posteriori (MAP) problem for abundance and endmember parameter estimation. A generalized expectation-maximization (EM) algorithm is used to minimize the MAP objective function. We tested the GMM approach on two real datasets, and showcased its efficacy for modeling endmember variability by comparing it to current popular methods.
{"title":"A Gaussian mixture model representation of endmember variability for spectral unmixing","authors":"Yuan Zhou, Anand Rangarajan, P. Gader","doi":"10.1109/WHISPERS.2016.8071802","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071802","url":null,"abstract":"Endmember variability complicates the problem of spectral unmixing. This variability is typically represented by probability distributions or spectral libraries. The present work describes a new distributional representation based on Gaussian Mixture Models (GMMs). The most common form in this setting is the Normal Compositional Model (NCM), wherein the endmembers for each pixel are modeled as samples drawn from unimodal Gaussians. In reality, however, the distribution of spectra from a material may be multi-modal. We first show that a linear combination of GMM random variables is also a GMM. This allows us to probabilistically formulate hyperspectral pixel likelihoods as combinations of independent endmember random variables. Then, after adding a reasonable smoothness and sparsity prior on the abundances, we obtain a standard Bayesian maximum a posteriori (MAP) problem for abundance and endmember parameter estimation. A generalized expectation-maximization (EM) algorithm is used to minimize the MAP objective function. We tested the GMM approach on two real datasets, and showcased its efficacy for modeling endmember variability by comparing it to current popular methods.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"239 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":"133796609","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.2016.8071795
A. Marinoni, H. Clenet
Typically, quantitative interpretation of Mars mineralogy from spectra can be retrieved by analyzing the overlaps of absorption features. It is possible to achieve a thorough description of the abundances of each mineral the considered scene is composed of by applying proper deconvolution techniques such as those based on modified Gaussian model (MGM). However, MGM-based methods are sensitive on initial parameters for statistical distribution definition, or they are very time consuming when fully automatized. In this paper, a new method for identification of minerals on Mars surface by means of higher order nonlinear hyperspectral unmixing framework is introduced. Abundance distribution of magmatic minerals (olivine and pyroxenes) compounds is retrieved according to polytope decomposition algorithm. Experimental results show how the proposed method is able to provide actual abundance maps which are highly correlated to those obtained by an automatized MGM-based technique.
{"title":"Identification of mafic minerals on Mars by nonlinear hyperspectral unmixing","authors":"A. Marinoni, H. Clenet","doi":"10.1109/WHISPERS.2016.8071795","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071795","url":null,"abstract":"Typically, quantitative interpretation of Mars mineralogy from spectra can be retrieved by analyzing the overlaps of absorption features. It is possible to achieve a thorough description of the abundances of each mineral the considered scene is composed of by applying proper deconvolution techniques such as those based on modified Gaussian model (MGM). However, MGM-based methods are sensitive on initial parameters for statistical distribution definition, or they are very time consuming when fully automatized. In this paper, a new method for identification of minerals on Mars surface by means of higher order nonlinear hyperspectral unmixing framework is introduced. Abundance distribution of magmatic minerals (olivine and pyroxenes) compounds is retrieved according to polytope decomposition algorithm. Experimental results show how the proposed method is able to provide actual abundance maps which are highly correlated to those obtained by an automatized MGM-based technique.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"6 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":"122093781","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.2016.8071768
Yuanchao Su, Xu Sun, Lianru Gao, Jun Yu Li, Bing Zhang
Endmember extraction is a key step in hyperspectral unmixing. This paper proposes a new endmember extraction framework based on the swarm intelligence algorithm. We adopt a discrete structure because pixels exist within a discrete frame. Traditional swarm intelligence algorithms produce stacked solutions based on similar endmembers in the same class. We introduce a “distance” factor into the objective function to limit the number of endmembers per class. We then propose three endmember extraction methods based on the artificial bee colony (ABC), ant colony optimization (ACO) and particle swarm optimization (PSO) algorithms. Experiments with both simulated and actual hyperspectral image data indicate that the proposed framework can significantly improve the accuracy of endmember extraction.
{"title":"Improved discrete swarm intelligence algorithms for endmember extraction in hyperspectral remote sensing image","authors":"Yuanchao Su, Xu Sun, Lianru Gao, Jun Yu Li, Bing Zhang","doi":"10.1109/WHISPERS.2016.8071768","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071768","url":null,"abstract":"Endmember extraction is a key step in hyperspectral unmixing. This paper proposes a new endmember extraction framework based on the swarm intelligence algorithm. We adopt a discrete structure because pixels exist within a discrete frame. Traditional swarm intelligence algorithms produce stacked solutions based on similar endmembers in the same class. We introduce a “distance” factor into the objective function to limit the number of endmembers per class. We then propose three endmember extraction methods based on the artificial bee colony (ABC), ant colony optimization (ACO) and particle swarm optimization (PSO) algorithms. Experiments with both simulated and actual hyperspectral image data indicate that the proposed framework can significantly improve the accuracy of endmember extraction.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"17 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":"124569590","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.2016.8071783
Shirin Hassanzadeh, A. Karami, Rob Heylen, P. Scheunders
In this paper, a new lossy compression method for hyperspectral images (HSI) is introduced based on the NonNegative Tucker Decomposition (NTD). HSI are considered as a 3D dataset: two spatial dimensions and one spectral dimension. The NTD algorithm decomposes the original data into a smaller 3D dataset (core tensor) and three matrices. In the proposed method, in order to find the optimal decomposition, the Block Coordinate Descent (BCD) method is used, which is initialized by using Compressed Sensing (CS). The obtained optimal core tensor and matrices are coded by applying arithmetic coding and finally the compressed dataset is transmitted. The proposed method is applied to real datasets. Our experimental results show that, in comparison with state-of-the-art lossy compression methods, the proposed method achieves the highest signal to noise ratio (SNR) at any desired compression ratio (CR) while noise reduction is simultaneously obtained.
{"title":"Compression of hyperspectral images using block coordinate descent search and compressed sensing","authors":"Shirin Hassanzadeh, A. Karami, Rob Heylen, P. Scheunders","doi":"10.1109/WHISPERS.2016.8071783","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071783","url":null,"abstract":"In this paper, a new lossy compression method for hyperspectral images (HSI) is introduced based on the NonNegative Tucker Decomposition (NTD). HSI are considered as a 3D dataset: two spatial dimensions and one spectral dimension. The NTD algorithm decomposes the original data into a smaller 3D dataset (core tensor) and three matrices. In the proposed method, in order to find the optimal decomposition, the Block Coordinate Descent (BCD) method is used, which is initialized by using Compressed Sensing (CS). The obtained optimal core tensor and matrices are coded by applying arithmetic coding and finally the compressed dataset is transmitted. The proposed method is applied to real datasets. Our experimental results show that, in comparison with state-of-the-art lossy compression methods, the proposed method achieves the highest signal to noise ratio (SNR) at any desired compression ratio (CR) while noise reduction is simultaneously obtained.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"39 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":"134538629","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.2016.8071778
Sujan Singh Niranjan, Neelima Chaube, J. Sarup
This paper presents the study of mineral and vegetation in explored fields around the San Juan coal mines west of Farmington, New Mexico. The purpose of this research work is to map the mineral rock & vegetation for statistically analyzing the study area. Pre-processing of Hyperspectral imagery (HSI) data is required for conversion from digital value to reflectance. Minimum Noise Fraction (MNF) and Pure Pixel Index (PPI) method is used for extraction of Endmember fraction. Spectral signature matching procedure is done with U.S. Geological Survey (USGS) Spectral library, which contain spectra of individual species that have been acquired at test sites representatives of varied terrain and climatic zones, observed in the field under natural conditions. Spectral Angle Mapper (SAM) technique is used for spectral analysis and mapping of image. Finally study area is mapped in two classes namely Carnallite mineral and Sagebrush vegetation plants. Land covered by Sagebrush plant is 8.31% and Carnallite is 1.41% of study area.
{"title":"Mapping of the carnallite mineral and sagebrush vegetation plant by using hyperspectral remote sensing and usgs spectral library","authors":"Sujan Singh Niranjan, Neelima Chaube, J. Sarup","doi":"10.1109/WHISPERS.2016.8071778","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071778","url":null,"abstract":"This paper presents the study of mineral and vegetation in explored fields around the San Juan coal mines west of Farmington, New Mexico. The purpose of this research work is to map the mineral rock & vegetation for statistically analyzing the study area. Pre-processing of Hyperspectral imagery (HSI) data is required for conversion from digital value to reflectance. Minimum Noise Fraction (MNF) and Pure Pixel Index (PPI) method is used for extraction of Endmember fraction. Spectral signature matching procedure is done with U.S. Geological Survey (USGS) Spectral library, which contain spectra of individual species that have been acquired at test sites representatives of varied terrain and climatic zones, observed in the field under natural conditions. Spectral Angle Mapper (SAM) technique is used for spectral analysis and mapping of image. Finally study area is mapped in two classes namely Carnallite mineral and Sagebrush vegetation plants. Land covered by Sagebrush plant is 8.31% and Carnallite is 1.41% of study area.","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":"129131889","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.2016.8071752
Jingliang Hu, Pedram Ghamisi, A. Schmitt, Xiaoxiang Zhu
In this paper, we propose an object-based fusion approach for the joint use of polarimetric synthetic aperture radar (PolSAR) and hyperspectral data. The proposed approach extracts information from both datasets based on an object-level, which is used here for land use classification. The achieved classification result infers that the proposed methodology improves the classification performance of both hyperspectral and PolSAR data and can properly gather complementary information of the two kinds of dataset. The fusion approach also considers that only limited training samples are available, which is often the case in remote sensing.
{"title":"Object based fusion of polarimetric SAR and hyperspectral imaging for land use classification","authors":"Jingliang Hu, Pedram Ghamisi, A. Schmitt, Xiaoxiang Zhu","doi":"10.1109/WHISPERS.2016.8071752","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071752","url":null,"abstract":"In this paper, we propose an object-based fusion approach for the joint use of polarimetric synthetic aperture radar (PolSAR) and hyperspectral data. The proposed approach extracts information from both datasets based on an object-level, which is used here for land use classification. The achieved classification result infers that the proposed methodology improves the classification performance of both hyperspectral and PolSAR data and can properly gather complementary information of the two kinds of dataset. The fusion approach also considers that only limited training samples are available, which is often the case in remote sensing.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"242 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":"133864817","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.2016.8071700
M. Fahad, Mingyi He, Yifan Zhang
There are two target detection algorithms which are commonly used in various applications. Both of them work on a related linear process, which makes them intensely related. This paper suggests a hyperspectral target detection algorithm which is a combination of CEM (Constrained Energy Minimization) and RXD (Reed-Xiaoli detector) algorithms to employ the advantages of both approaches to improve detection performance. The comparison of different target detection algorithms are performed by Receiver Operating Characteristic (ROC) Curves. The experimental result shows that this combination can efficiently improves the detection performance.
{"title":"Combination of CEM & RXD for target detection in hyperspectral images","authors":"M. Fahad, Mingyi He, Yifan Zhang","doi":"10.1109/WHISPERS.2016.8071700","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071700","url":null,"abstract":"There are two target detection algorithms which are commonly used in various applications. Both of them work on a related linear process, which makes them intensely related. This paper suggests a hyperspectral target detection algorithm which is a combination of CEM (Constrained Energy Minimization) and RXD (Reed-Xiaoli detector) algorithms to employ the advantages of both approaches to improve detection performance. The comparison of different target detection algorithms are performed by Receiver Operating Characteristic (ROC) Curves. The experimental result shows that this combination can efficiently improves the detection performance.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"10 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":"121471264","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.2016.8071656
Pedram Ghamisi, R. Souza, J. Benediktsson, Xiaoxiang Zhu, L. Rittner, R. Lotufo
In this paper, a novel approach is proposed for the spectral-spatial classification of hyperspectral images. The proposed classification approach is based on a novel filtering technique, here entitled as extended extinction profile (EEP). The proposed classification approach is applied on two well-known data sets: Pavia University and Indian Pines; and the obtained results have been compared with one of the strongest filtering approaches in the literature named extended attribute profile (EAP). Results confirm that the proposed approach is able to efficiently extract spatial information for the classification of hyperspectral images.
{"title":"Extended extinction profile for the classification of hyperspectral images","authors":"Pedram Ghamisi, R. Souza, J. Benediktsson, Xiaoxiang Zhu, L. Rittner, R. Lotufo","doi":"10.1109/WHISPERS.2016.8071656","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071656","url":null,"abstract":"In this paper, a novel approach is proposed for the spectral-spatial classification of hyperspectral images. The proposed classification approach is based on a novel filtering technique, here entitled as extended extinction profile (EEP). The proposed classification approach is applied on two well-known data sets: Pavia University and Indian Pines; and the obtained results have been compared with one of the strongest filtering approaches in the literature named extended attribute profile (EAP). Results confirm that the proposed approach is able to efficiently extract spatial information for the classification of hyperspectral images.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"22 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":"121098136","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.2016.8071690
R. Roscher, J. Behmann, Anne-Katrin Mahlein, L. Plümer
We analyze the benefit of using topographic dictionaries for a sparse representation (SR) approach for the detection of Cercospora leaf spot disease symptoms on sugar beet plants. Topographic dictionaries are an arranged set of basis elements in which neighbored dictionary elements tend to cause similar activations in the SR approach. In this paper, the dictionary is obtained from samples of a healthy plant and partly build in a topographic way by using hyperspectral as well as geometry information, i.e. depth and inclination. It turns out that hyperspectral signals of leafs show a typical structure depending on depth and inclination and thus, both influences can be disentangled in our approach. Rare signals which do not fit into this model, e.g. leaf veins, are also captured in the dictionary in a non-topographic way. A reconstruction error index is used as indicator, in which disease symptoms can be distinguished from healthy plant regions. The advantage of the presented approach is that full spectral and geometry information is needed only once to built the dictionary, whereas the sparse reconstruction is done solely on hyperspectral information.
{"title":"On the benefit of topographic dictionaries for detecting disease symptoms on hyperspectral 3D plant models","authors":"R. Roscher, J. Behmann, Anne-Katrin Mahlein, L. Plümer","doi":"10.1109/WHISPERS.2016.8071690","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071690","url":null,"abstract":"We analyze the benefit of using topographic dictionaries for a sparse representation (SR) approach for the detection of Cercospora leaf spot disease symptoms on sugar beet plants. Topographic dictionaries are an arranged set of basis elements in which neighbored dictionary elements tend to cause similar activations in the SR approach. In this paper, the dictionary is obtained from samples of a healthy plant and partly build in a topographic way by using hyperspectral as well as geometry information, i.e. depth and inclination. It turns out that hyperspectral signals of leafs show a typical structure depending on depth and inclination and thus, both influences can be disentangled in our approach. Rare signals which do not fit into this model, e.g. leaf veins, are also captured in the dictionary in a non-topographic way. A reconstruction error index is used as indicator, in which disease symptoms can be distinguished from healthy plant regions. The advantage of the presented approach is that full spectral and geometry information is needed only once to built the dictionary, whereas the sparse reconstruction is done solely on hyperspectral information.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"21 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":"115688389","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.2016.8071663
P. V. Arun, K. Buddhiraju, A. Porwal
In this paper, we investigate the use of coarse image features for predicting class label distributions at a finer scale. The major contributions of this work are 1) use of coarse image features to improve the optimization formulation of conventional rank based approaches 2) use of inter class compatibility information from coarse images to refine the predicted target distribution 3) an enhanced unsupervised variogram based sub-pixel mapping approach 4) inclusion of abundance estimation uncertainty in the unmixing process. The proposed modifications on rank based and variogram based approaches have produced an accuracy improvement of 10–15%. The sensitivities of these approaches towards tunable parameters are also analyzed.
{"title":"Integration of contextual knowledge in unsupervised sub-pixel classification","authors":"P. V. Arun, K. Buddhiraju, A. Porwal","doi":"10.1109/WHISPERS.2016.8071663","DOIUrl":"https://doi.org/10.1109/WHISPERS.2016.8071663","url":null,"abstract":"In this paper, we investigate the use of coarse image features for predicting class label distributions at a finer scale. The major contributions of this work are 1) use of coarse image features to improve the optimization formulation of conventional rank based approaches 2) use of inter class compatibility information from coarse images to refine the predicted target distribution 3) an enhanced unsupervised variogram based sub-pixel mapping approach 4) inclusion of abundance estimation uncertainty in the unmixing process. The proposed modifications on rank based and variogram based approaches have produced an accuracy improvement of 10–15%. The sensitivities of these approaches towards tunable parameters are also analyzed.","PeriodicalId":369281,"journal":{"name":"2016 8th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)","volume":"17 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":"116567482","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：