Nowadays, diverse sensor technologies allow us to
measure different aspects of objects on the Ear... more Nowadays, diverse sensor technologies allow us to measure different aspects of objects on the Earth surface (spectral characteristics in hyperspectral (HS) images, height in Light Detection And Ranging (LiDAR) data, etc), with increasing spectral and spatial resolutions. Remote sensing images of very high geometrical resolution can provide a precise and detailed representation of the monitored scene. Thus, the spatial infor- mation is fundamental for many applications. Morphological profiles (MPs) and attribute profiles (APs) have been widely used to model the spatial information of very high resolution (VHR) remote sensing images. MPs are obtained by comput- ing a sequence of morphological operators based on geodesic reconstruction. However, both morphological operators based on geodesic reconstruction and attribute filters are connected filters, and hence suffer the problem of ‘leakage’ (i.e., regions related to different structures in the image that happen to be connected by spurious links will be considered as a single object). Objects expected to disappear at a given stage still remain present when they are connected with other objects in the image. As a consequence, the attributes of small objects will be mixed with their larger connected objects, leading to poor performances on post-applications (e.g., classification). In this work, we introduce morphological partial reconstruction for spatial information modeling of VHR urban remote sensing images. The ultimate goal of partial reconstruction is to extract spatial features which better model the attributes of different objects leading to improved classification performances. These methods are applied to three datasets with different sensor modalities, resolutions and properties (including panchromatic, hyperspectral and LiDAR images), and their effectiveness and robustness are quantitatively and qualitatively evaluated. In addition, morphological partial reconstruction codes introduced in this paper have been implemented in a MATLAB toolbox http://telin.ugent.be/~wliao/Partial Reconstruction that is made available to the community.
2009 Ninth International Conference on Intelligent Systems Design and Applications, 2009
ABSTRACT The theory of Chinese character intelligent formation considers that Chinese characters ... more ABSTRACT The theory of Chinese character intelligent formation considers that Chinese characters are formed by components according to character structure; all the components are the topological mapping of basic elements in the character structure. The mapping method of basic elements in different character structures is one of the key technologies. This paper carried on a thorough analysis to the transformation of basic elements, proposed the topological mapping method based on affine transformation. 27533 Chinese characters in GB18030-2000 standard were taken as experiment subject, a platform for Chinese character intelligent formation system was developed and all the characters were formed in the platform.
Hyperspectral imagery contains a wealth of spectral and spatial information that can improve targ... more Hyperspectral imagery contains a wealth of spectral and spatial information that can improve target detection and recognition performance. Typically, spectral information is inferred pixel-based, while spatial information related to texture, context and geometry are deduced on a per-object basis. Existing feature extraction methods cannot fully utilize both the spectral and spatial information. Data fusion by simply stacking different feature sources together does not take into account the differences between feature sources. In this paper, we propose a feature fusion method to couple dimension reduction and data fusion of the pixel- and object-based features of hyperspectral imagery. The proposed method takes into account the properties of different feature sources, and makes full advantage of both the pixel- and object-based features through the fusion graph. Experimental results on classification of urban hyperspectral remote sensing image are very encouraging.
Nowadays, advanced technology in remote sensing allows us to acquire multi-sensor and multi-resol... more Nowadays, advanced technology in remote sensing allows us to acquire multi-sensor and multi-resolution data from the same geographic region. Fusion of these data sources for classification purposes however remains challenging. We propose a novel framework for fusion of low spatial resolution Thermal Infrared (TI) hyperspectral (HS) and high spatial resolution RGB data. First, we perform image fusion on the TIHS image by using visible RGB image and guided filtering in PCA (Principal Component Analysis) domain. Then, we couple feature extraction and data fusion of spectral features (from TI HS data) and spatial features (morphological features generated on RGB image) through a supervised fusion graph. Finally, the fused features are used by a SVM (Support Vector Machine) classifier to generate the final classification map. Experimental results on the classification of fusing real TI HS and RGB images demonstrate the effectiveness of the proposed method both visually and quantitatively.
2008 Congress on Image and Signal Processing, 2008
ABSTRACT A new feature extraction approach based on kernel independent component analysis (Kernel... more ABSTRACT A new feature extraction approach based on kernel independent component analysis (Kernel ICA) is proposed in this paper. The Kernel ICA is applied to learn basis vector for feature extraction, and then the basis vector is used as a template model to extract the edge feature from the testing images which are completely different from the training image. The simulating experiment shows that the approach proposed in this paper has a better performance than ICA.
Abstract When using morphological features for the classification of high resolution hyperspectra... more Abstract When using morphological features for the classification of high resolution hyperspectral images from urban areas, one should consider two important issues. The first one is that classical morphological openings and closings degrade the object boundaries and deform the object shapes. Morphological openings and closings by reconstruction can avoid this problem, but this process leads to some undesirable effects. Objects expected to disappear at a certain scale remain present when using morphological openings and ...
Nowadays, diverse sensor technologies allow us to
measure different aspects of objects on the Ear... more Nowadays, diverse sensor technologies allow us to measure different aspects of objects on the Earth surface (spectral characteristics in hyperspectral (HS) images, height in Light Detection And Ranging (LiDAR) data, etc), with increasing spectral and spatial resolutions. Remote sensing images of very high geometrical resolution can provide a precise and detailed representation of the monitored scene. Thus, the spatial infor- mation is fundamental for many applications. Morphological profiles (MPs) and attribute profiles (APs) have been widely used to model the spatial information of very high resolution (VHR) remote sensing images. MPs are obtained by comput- ing a sequence of morphological operators based on geodesic reconstruction. However, both morphological operators based on geodesic reconstruction and attribute filters are connected filters, and hence suffer the problem of ‘leakage’ (i.e., regions related to different structures in the image that happen to be connected by spurious links will be considered as a single object). Objects expected to disappear at a given stage still remain present when they are connected with other objects in the image. As a consequence, the attributes of small objects will be mixed with their larger connected objects, leading to poor performances on post-applications (e.g., classification). In this work, we introduce morphological partial reconstruction for spatial information modeling of VHR urban remote sensing images. The ultimate goal of partial reconstruction is to extract spatial features which better model the attributes of different objects leading to improved classification performances. These methods are applied to three datasets with different sensor modalities, resolutions and properties (including panchromatic, hyperspectral and LiDAR images), and their effectiveness and robustness are quantitatively and qualitatively evaluated. In addition, morphological partial reconstruction codes introduced in this paper have been implemented in a MATLAB toolbox http://telin.ugent.be/~wliao/Partial Reconstruction that is made available to the community.
2009 Ninth International Conference on Intelligent Systems Design and Applications, 2009
ABSTRACT The theory of Chinese character intelligent formation considers that Chinese characters ... more ABSTRACT The theory of Chinese character intelligent formation considers that Chinese characters are formed by components according to character structure; all the components are the topological mapping of basic elements in the character structure. The mapping method of basic elements in different character structures is one of the key technologies. This paper carried on a thorough analysis to the transformation of basic elements, proposed the topological mapping method based on affine transformation. 27533 Chinese characters in GB18030-2000 standard were taken as experiment subject, a platform for Chinese character intelligent formation system was developed and all the characters were formed in the platform.
Hyperspectral imagery contains a wealth of spectral and spatial information that can improve targ... more Hyperspectral imagery contains a wealth of spectral and spatial information that can improve target detection and recognition performance. Typically, spectral information is inferred pixel-based, while spatial information related to texture, context and geometry are deduced on a per-object basis. Existing feature extraction methods cannot fully utilize both the spectral and spatial information. Data fusion by simply stacking different feature sources together does not take into account the differences between feature sources. In this paper, we propose a feature fusion method to couple dimension reduction and data fusion of the pixel- and object-based features of hyperspectral imagery. The proposed method takes into account the properties of different feature sources, and makes full advantage of both the pixel- and object-based features through the fusion graph. Experimental results on classification of urban hyperspectral remote sensing image are very encouraging.
Nowadays, advanced technology in remote sensing allows us to acquire multi-sensor and multi-resol... more Nowadays, advanced technology in remote sensing allows us to acquire multi-sensor and multi-resolution data from the same geographic region. Fusion of these data sources for classification purposes however remains challenging. We propose a novel framework for fusion of low spatial resolution Thermal Infrared (TI) hyperspectral (HS) and high spatial resolution RGB data. First, we perform image fusion on the TIHS image by using visible RGB image and guided filtering in PCA (Principal Component Analysis) domain. Then, we couple feature extraction and data fusion of spectral features (from TI HS data) and spatial features (morphological features generated on RGB image) through a supervised fusion graph. Finally, the fused features are used by a SVM (Support Vector Machine) classifier to generate the final classification map. Experimental results on the classification of fusing real TI HS and RGB images demonstrate the effectiveness of the proposed method both visually and quantitatively.
2008 Congress on Image and Signal Processing, 2008
ABSTRACT A new feature extraction approach based on kernel independent component analysis (Kernel... more ABSTRACT A new feature extraction approach based on kernel independent component analysis (Kernel ICA) is proposed in this paper. The Kernel ICA is applied to learn basis vector for feature extraction, and then the basis vector is used as a template model to extract the edge feature from the testing images which are completely different from the training image. The simulating experiment shows that the approach proposed in this paper has a better performance than ICA.
Abstract When using morphological features for the classification of high resolution hyperspectra... more Abstract When using morphological features for the classification of high resolution hyperspectral images from urban areas, one should consider two important issues. The first one is that classical morphological openings and closings degrade the object boundaries and deform the object shapes. Morphological openings and closings by reconstruction can avoid this problem, but this process leads to some undesirable effects. Objects expected to disappear at a certain scale remain present when using morphological openings and ...
Uploads
Papers by Wenzhi Liao
measure different aspects of objects on the Earth surface (spectral
characteristics in hyperspectral (HS) images, height in Light
Detection And Ranging (LiDAR) data, etc), with increasing
spectral and spatial resolutions. Remote sensing images of very
high geometrical resolution can provide a precise and detailed
representation of the monitored scene. Thus, the spatial infor-
mation is fundamental for many applications. Morphological
profiles (MPs) and attribute profiles (APs) have been widely
used to model the spatial information of very high resolution
(VHR) remote sensing images. MPs are obtained by comput-
ing a sequence of morphological operators based on geodesic
reconstruction. However, both morphological operators based
on geodesic reconstruction and attribute filters are connected
filters, and hence suffer the problem of ‘leakage’ (i.e., regions
related to different structures in the image that happen to
be connected by spurious links will be considered as a single
object). Objects expected to disappear at a given stage still
remain present when they are connected with other objects in
the image. As a consequence, the attributes of small objects
will be mixed with their larger connected objects, leading to
poor performances on post-applications (e.g., classification). In
this work, we introduce morphological partial reconstruction
for spatial information modeling of VHR urban remote sensing
images. The ultimate goal of partial reconstruction is to extract
spatial features which better model the attributes of different
objects leading to improved classification performances. These
methods are applied to three datasets with different sensor modalities, resolutions and properties (including panchromatic,
hyperspectral and LiDAR images), and their effectiveness and
robustness are quantitatively and qualitatively evaluated. In
addition, morphological partial reconstruction codes introduced
in this paper have been implemented in a MATLAB toolbox
http://telin.ugent.be/~wliao/Partial Reconstruction that is made
available to the community.
measure different aspects of objects on the Earth surface (spectral
characteristics in hyperspectral (HS) images, height in Light
Detection And Ranging (LiDAR) data, etc), with increasing
spectral and spatial resolutions. Remote sensing images of very
high geometrical resolution can provide a precise and detailed
representation of the monitored scene. Thus, the spatial infor-
mation is fundamental for many applications. Morphological
profiles (MPs) and attribute profiles (APs) have been widely
used to model the spatial information of very high resolution
(VHR) remote sensing images. MPs are obtained by comput-
ing a sequence of morphological operators based on geodesic
reconstruction. However, both morphological operators based
on geodesic reconstruction and attribute filters are connected
filters, and hence suffer the problem of ‘leakage’ (i.e., regions
related to different structures in the image that happen to
be connected by spurious links will be considered as a single
object). Objects expected to disappear at a given stage still
remain present when they are connected with other objects in
the image. As a consequence, the attributes of small objects
will be mixed with their larger connected objects, leading to
poor performances on post-applications (e.g., classification). In
this work, we introduce morphological partial reconstruction
for spatial information modeling of VHR urban remote sensing
images. The ultimate goal of partial reconstruction is to extract
spatial features which better model the attributes of different
objects leading to improved classification performances. These
methods are applied to three datasets with different sensor modalities, resolutions and properties (including panchromatic,
hyperspectral and LiDAR images), and their effectiveness and
robustness are quantitatively and qualitatively evaluated. In
addition, morphological partial reconstruction codes introduced
in this paper have been implemented in a MATLAB toolbox
http://telin.ugent.be/~wliao/Partial Reconstruction that is made
available to the community.