Stacked Autoencoders for Multiclass Change Detection in Hyperspectral Images

Lopez-Fandino, Javier; Garea, Alberto S.; Heras, Dora B.; Argüello, Francisco

doi:10.1109/igarss.2018.8518338

Cited by 45 publications

(34 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We validate the proposed method on two bi-temporal scenes (López-Fandiño et al, 2018) We compared the proposed method to three unsupervised methods:…”

Section: Resultsmentioning

confidence: 99%

Deep No Learning Approach for Unsupervised Change Detection in Hyperspectral Images

Saha

Kondmann

Zhu

2021

ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

Abstract. Unsupervised deep transfer-learning based change detection (CD) methods require pre-trained feature extractor that can be used to extract semantic features from the target bi-temporal scene. However, it is difficult to obtain such feature extractors for hyperspectral images. Moreover, it is not trivial to reuse the models trained with the multispectral images for the hyperspectral images due to the significant difference in number of spectral bands. While hyperspectral images show large number of spectral bands, they generally show much less spatial complexity, thus reducing the requirement of large receptive fields of convolution filters. Recent works in the computer vision have shown that even untrained networks can yield remarkable result in different tasks like super-resolution and surface reconstruction. Motivated by this, we make a bold proposition that untrained deep model, initialized with some weight initialization strategy can be used to extract useful semantic features from bi-temporal hyperspectral images. Thus, we couple an untrained network with Deep Change Vector Analysis (DCVA), a popular method for unsupervised CD, to propose an unsupervised CD method for hyperspectral images. We conduct experiments on two hyperspectral CD data sets, and the results demonstrate advantages of the proposed unsupervised method over other competitors.

show abstract

“…We validate the proposed method on two bi-temporal scenes (López-Fandiño et al, 2018) We compared the proposed method to three unsupervised methods:…”

Section: Resultsmentioning

confidence: 99%

Deep No Learning Approach for Unsupervised Change Detection in Hyperspectral Images

Saha

Kondmann

Zhu

2021

ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

show abstract

“…Some datasets with larger image sizes, say, width/height of a few thousand pixels, are limited with the number of annotated images or categories [87], [136], [147], [153], [157], [158]. Furthermore, quite a few datasets contain one or several images, especially those for semantic segmentation [125], [133], [134] and change detection [97], [145]- [151], [155], [156], [160], which are limited by the high cost of pixel-wise annotation. As a result, the scale limitations in size and number of images could easily lead to performance saturation for interpretation algorithms.…”

Section: ) Dataset Scalementioning

confidence: 99%

On Creating Benchmark Dataset for Aerial Image Interpretation: Reviews, Guidances, and Million-AID

Yang

Xia

et al. 2021

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

103

View full text Add to dashboard Cite

The past years have witnessed great progress on remote sensing (RS) image interpretation and its wide applications. With RS images becoming more accessible than ever before, there is an increasing demand for the automatic interpretation of these images. In this context, the benchmark datasets serve as essential prerequisites for developing and testing intelligent interpretation algorithms. After reviewing existing benchmark datasets in the research community of RS image interpretation, this article discusses the problem of how to efficiently prepare a suitable benchmark dataset for RS image interpretation. Specifically, we first analyze the current challenges of developing intelligent algorithms for RS image interpretation with bibliometric investigations. We then present the general guidances on creating benchmark datasets in efficient manners. Following the presented guidances, we also provide an example on building RS image dataset, i.e., Million-AID 1 , a new large-scale benchmark dataset containing a million instances for RS image scene classification. Several challenges and perspectives in RS image annotation are finally discussed to facilitate the research in benchmark dataset construction. We do hope this paper will provide the RS community an overall perspective on constructing large-scale and practical image datasets for further research, especially datadriven ones.

show abstract

“…At present, the researches of hyperspectral images for change detection are not substantial, mainly due to the difficulty of image preprocessing, annotation, and end-element decomposition. We have collected two relevant datasets as examples, the Hyperspectral Change Detection Dataset [54] and Hyperspectral image (HSI) Datasets [15], as shown in Figure 7. Based on the current RS technology, the hyperspectral images usually possess high spectral resolution but low spatial resolution.…”

Section: Hyperspectral Imagesmentioning

confidence: 99%

“…To some extent, AE reduces the strict requirements of the layer parameters. Similar to DBN, stacked autocoder (SAE) [54] is formed by the accumulation of AE neurons [119]. Derived from AE structure, stacked contractual autocoder (SCAE) is used for feature extraction and noise suppression in iterative encode decode structure [120].…”

Section: Naive Dnnmentioning

confidence: 99%

A Survey of Change Detection Methods Based on Remote Sensing Images for Multi-Source and Multi-Objective Scenarios

2020

View full text Add to dashboard Cite

Quantities of multi-temporal remote sensing (RS) images create favorable conditions for exploring the urban change in the long term. However, diverse multi-source features and change patterns bring challenges to the change detection in urban cases. In order to sort out the development venation of urban change detection, we make an observation of the literatures on change detection in the last five years, which focuses on the disparate multi-source RS images and multi-objective scenarios determined according to scene category. Based on the survey, a general change detection framework, including change information extraction, data fusion, and analysis of multi-objective scenarios modules, is summarized. Owing to the attributes of input RS images affect the technical selection of each module, data characteristics and application domains across different categories of RS images are discussed firstly. On this basis, not only the evolution process and relationship of the representative solutions are elaborated in the module description, through emphasizing the feasibility of fusing diverse data and the manifold application scenarios, we also advocate a complete change detection pipeline. At the end of the paper, we conclude the current development situation and put forward possible research direction of urban change detection, in the hope of providing insights to the following research.

show abstract

Stacked Autoencoders for Multiclass Change Detection in Hyperspectral Images

Cited by 45 publications

References 6 publications

Deep No Learning Approach for Unsupervised Change Detection in Hyperspectral Images

Deep No Learning Approach for Unsupervised Change Detection in Hyperspectral Images

On Creating Benchmark Dataset for Aerial Image Interpretation: Reviews, Guidances, and Million-AID

A Survey of Change Detection Methods Based on Remote Sensing Images for Multi-Source and Multi-Objective Scenarios

Contact Info

Product

Resources

About