Series Of Multispectral Images Research Articles

Resting-state EEG signal classification of amnestic mild cognitive impairment with type 2 diabetes mellitus based on multispectral image and convolutional neural network

Objective. The purpose of this study is to judge whether this combination method of multispectral image and convolutional neural network (CNN) method can be used to distinguish amnestic mild cognitive impairment (aMCI) with Type 2 diabetes mellitus (T2DM) and normal controls (NC) with T2DM effectively. Approach. In this study, the authors first combined EEG signals from aMCI patients with T2DM and NC with T2DM on five different frequency bands, including Theta, Alpha1, Alpha2, Beta1, and Beta2. Then, the authors converted these time series into a series of multispectral images. Finally, the images data were classified with the CNN method. Main results. The classification effects of up to 89%, 91%, and 92% are obtained on the three combinations of frequency bands: Theta, Alpha1, and Alpha2; Alpha1, Alpha2, and Beta1; and Alpha2, Beta1, and Beta2. The spatial properties of EEG signals are highlighted, and its classification performance is found to be better than all the previous methods in the field of aMCI and T2DM diagnosis. The combination of multispectral images and CNN can be used as an effective biomarker for distinguishing the EEG signals in patients with aMCI and T2DM and in patients with NC with T2DM. Significance. The combined approach used in this paper provides a new perspective for the analysis of EEG signals in patients with aMCI and T2DM.

CHANGE DETECTION WITH MULTI-SOURCE DEFECTIVE REMOTE SENSING IMAGES BASED ON EVIDENTIAL FUSION

Remote sensing images with clouds, shadows or stripes are usually considered as defective data which limit their application for change detection. This paper proposes a method to fuse a series of defective images as evidences for change detection. In the proposed method, post-classification comparison process is firstly performed on multi-source defective images. Then, the classification results of all the images, together with their corresponding confusion matrixes are used to calculate the Basic Belief Assignment (BBA) of each pixel. Further, based on the principle of Dempster-Shafer evidence theory, a BBA redistribution process is introduced to deal with the defective parts of multi-source data. At last, evidential fusion and decision making rules are applied on the pixel level, and the final map of change detection can be derived. The proposed method can finish change detection with data fusion and image completion in one integrated process, which makes use of the complementary and redundant information from the input images. The method is applied to a case study of landslide barrier lake formed in Aug. 3rd, 2014, with a series of multispectral images from different sensors of GF-1 satellite. Result shows that the proposed method can not only complete the defective parts of the input images, but also provide better change detection accuracy than post-classification comparison method with single pair of pre- and post-change images. Subsequent analysis indicates that high conflict degree between evidences is the main source of errors in the result. Finally, some possible reasons that result in evidence conflict on the pixel level are analysed.

CHANGE DETECTION WITH MULTI-SOURCE DEFECTIVE REMOTE SENSING IMAGES BASED ON EVIDENTIAL FUSION

Abstract. Remote sensing images with clouds, shadows or stripes are usually considered as defective data which limit their application for change detection. This paper proposes a method to fuse a series of defective images as evidences for change detection. In the proposed method, post-classification comparison process is firstly performed on multi-source defective images. Then, the classification results of all the images, together with their corresponding confusion matrixes are used to calculate the Basic Belief Assignment (BBA) of each pixel. Further, based on the principle of Dempster-Shafer evidence theory, a BBA redistribution process is introduced to deal with the defective parts of multi-source data. At last, evidential fusion and decision making rules are applied on the pixel level, and the final map of change detection can be derived. The proposed method can finish change detection with data fusion and image completion in one integrated process, which makes use of the complementary and redundant information from the input images. The method is applied to a case study of landslide barrier lake formed in Aug. 3rd, 2014, with a series of multispectral images from different sensors of GF-1 satellite. Result shows that the proposed method can not only complete the defective parts of the input images, but also provide better change detection accuracy than post-classification comparison method with single pair of pre- and post-change images. Subsequent analysis indicates that high conflict degree between evidences is the main source of errors in the result. Finally, some possible reasons that result in evidence conflict on the pixel level are analysed.

Multitemporal Fusion for the Detection of Static Spatial Patterns in Multispectral Satellite Images—With Application to Archaeological Survey

We evaluate and further develop a multitemporal fusion strategy that we use to detect the location of ancient settlement sites in the Near East and to map their distribution, a spatial pattern that remains static over time. For each ASTER images that has been acquired in our survey area in northeastern Syria, we use a pattern classification strategy to map locations with a multispectral signal similar to the one from (few) known archaeological sites nearby. We obtain maps indicating the presence of anthrosol-soils that formed in the location of ancient settlements and that have a distinct spectral pattern under certain environmental conditions-and find that pooling the probability maps from all available time points reduces the variance of the spatial anthrosol pattern significantly. Removing biased classification maps-i.e., those that rank last when comparing the probability maps with the (limited) ground truth we have- reduces the overall prediction error even further, and we estimate optimal weights for each image using a nonnegative least squares regression strategy. The ranking and pooling strategy approach we propose in this study shows a significant improvement over the plain averaging of anthrosol probability maps that we used in an earlier attempt to map archaeological sites in a 20 000-km 2 area in northern Mesopotamia, and we expect it to work well in other surveying tasks that aim in mapping static surface patterns with limited ground truth in long series of multispectral images.

Multispectral image compression based on DSC combined with CCSDS-IDC.

Remote sensing multispectral image compression encoder requires low complexity, high robust, and high performance because it usually works on the satellite where the resources, such as power, memory, and processing capacity, are limited. For multispectral images, the compression algorithms based on 3D transform (like 3D DWT, 3D DCT) are too complex to be implemented in space mission. In this paper, we proposed a compression algorithm based on distributed source coding (DSC) combined with image data compression (IDC) approach recommended by CCSDS for multispectral images, which has low complexity, high robust, and high performance. First, each band is sparsely represented by DWT to obtain wavelet coefficients. Then, the wavelet coefficients are encoded by bit plane encoder (BPE). Finally, the BPE is merged to the DSC strategy of Slepian-Wolf (SW) based on QC-LDPC by deep coupling way to remove the residual redundancy between the adjacent bands. A series of multispectral images is used to test our algorithm. Experimental results show that the proposed DSC combined with the CCSDS-IDC (DSC-CCSDS)-based algorithm has better compression performance than the traditional compression approaches.

High-fidelity image reproduction using angular distribution of reflected spectral intensity

A method to reproduce images of an object under various observation conditions is presented. In this method, a series of multispectral images is captured by rotating the object under a point light source of which spectral power distribution and the position are known. The captured images are decomposed into diffuse and specular reflection component images based on the dichromatic reflection and the Lambertian models. Next, the incident angle of the illumination light and the angle between viewing direction and regular reflection are calculated based on observation geometry. Finally, the image under observation geometry is synthesized using the light-ray rearrangement technique. The experiments are carried out using two-dimensional objects, leather and fabric. Most of the synthesized images are shown to be the same as the images actually captured under the assumed illumination geometry, even if the object has complex reflection like fabric for which it is difficult to apply the reflection model used in computer graphics.