A large number of publications have incorporated deep learning in the process of remote sensing change detection. In these Deep Learning Change Detection (DLCD) publications, deep learning methods have demonstrated their superiority over conventional change detection methods. However, the theoretical underpinnings of why deep learning improves the performance of change detection remain unresolved. As of today, few in-depth reviews have investigated the mechanisms of DLCD. Without such a review, five critical questions remain unclear. Does DLCD provide improved information representation for change detection? If so, how? How to select an appropriate DLCD method and why? How much does each type of change benefits from DLCD in terms of its performance? What are the major limitations of existing DLCD methods and what are the prospects for DLCD? To address these five questions, we reviewed according to the following strategies. We grouped the DLCD information assemblages into the four unique dimensions of remote sensing: spectral, spatial, temporal, and multi-sensor. For the extraction of information in each dimension, the difference between DLCD and conventional change detection methods was compared. We proposed a taxonomy of existing DLCD methods by dividing them into two distinctive pools: separate and coupled models. Their advantages, limitations, applicability, and performance were thoroughly investigated and explicitly presented. We examined the variations in performance between DLCD and conventional change detection. We depicted two limitations of DLCD, i.e. training sample and hardware and software dilemmas. Based on these analyses, we identified directions for future developments. As a result of our review, we found that DLCD’s advantages over conventional change detection can be attributed to three factors: improved information representation; improved change detection methods; and performance enhancements. DLCD has to surpass the limitations with regard to training samples and computing infrastructure. We envision this review can boost developments of deep learning in change detection applications.
Effort has been devoted to automatic or semi-automatic individual tree detection and crown delineation from remotely sensed data, aiming at the effective creation of forest inventory, to assist forest management and ecosystem modelling. However, there is no uniform evaluation scheme to assess the resulting products. This paper reviews the available techniques and essential considerations for evaluating remote sensing-based forest inventory products, including detected individual tree locations, crown delineation maps, forest structure parameters, and species classification results. The reference data collection, individual tree matching method, mismatched individual tree crown treatment, and stratification criteria are summarized as four key factors in the evaluation process. Additionally, a discussion of future working directions, regarding limitations of existing techniques, multi-scale evaluation, sampling schemes, uncertainty analysis, and results interpretation, is also included. A new accuracy assessment framework is proposed, and an accuracy report example is provided.
13 July 2016 in International Journal of Remote Sensing
Over the past decade, the incorporation of spatial information has drawn increasing attention in multispectral and hyperspectral data analysis. In particular, the property of spatial autocorrelation among pixels has shown great potential for improving understanding of remotely sensed imagery. In this paper, we provide a comprehensive review of the state-of-the-art techniques in incorporating spatial information in image classification and spectral unmixing. For image classification, spatial information is accounted for in the stages of pre-classification, sample selection, classifiers, post-classification, and accuracy assessment. With regards to spectral unmixing, spatial information is discussed in the context of endmember extraction, selection of endmember combinations, and abundance estimation. Finally, a perspective on future research directions for advancing spatial-spectral methods is offered.
14 June 2013 in International Journal of Remote Sensing
Knowledge of snow cover is essential to understanding the global water and energy cycle. Thresholding the normalized difference snow index (NDSI) image is a method frequently used to map snow cover from remotely sensed data. However, the threshold is dependent on the scenario and needs to be determined accordingly. In this study, nine automatic thresholding methods were tested on the NDSI. Comparisons of the automatic thresholding methods, optimal threshold, and support vector machine (SVM) classification show that Otsu's and Nie's methods appear to be the most robust among the nine automatic thresholding methods, achieving comparable accuracies with the latter two approaches. In addition, NDSI from the digital number (DN) can be an efficient substitution for NDSI obtained from atmospherically or topographically corrected data, with similar accuracy.