Abstract:Various Earth anomalies have destroyed the stable, balanced state, resulting in fatalities and serious destruction of property. With the advantages of large-scale and precise observation, high-resolution remote sensing images have been widely used for anomaly monitoring and localization. Powered by the deep representation, the existing methods have achieved remarkable advances, primarily in classification and change detection techniques. However, labeled samples are difficult to acquire due to the low probability of anomaly occurrence, and the trained models are limited to fixed anomaly categories, which hinders the application for anomalies with few samples or unknown anomalies. In this paper, to tackle this problem, we propose the anomaly change detection (AnomalyCD) technique, which accepts time-series observations and learns to identify anomalous changes by learning from the historical normal change pattern. Compared to the existing techniques, AnomalyCD processes an unfixed number of time steps and can localize the various anomalies in a unified manner, without human supervision. To benchmark AnomalyCD, we constructed a high-resolution dataset with time-series images dedicated to various Earth anomalies (the AnomalyCDD dataset). AnomalyCDD contains high-resolution (from 0.15 to 2.39 m/pixel), time-series (from 3 to 7 time steps), and large-scale images (1927.93 km2 in total) collected globally Furthermore, we developed a zero-shot baseline model (AnomalyCDM), which implements the AnomalyCD technique by extracting a general representation from the segment anything model (SAM) and conducting temporal comparison to distinguish the anomalous changes from normal changes. AnomalyCDM is designed as a two-stage workflow to enhance the efficiency, and has the ability to process the unseen images directly, without retraining for each scene.

Kernel Anomalous Change Detection for Remote Sensing Imagery

AnomalyCD: A Benchmark for Earth Anomaly Change Detection with High-Resolution and Time-Series Observations

Nonlinear Cook distance for Anomalous Change Detection

Nonlinear Anomaly Detection Based on Spectral–Spatial Composite Kernel for Hyperspectral Images

Unsupervised Anomaly and Change Detection with Multivariate Gaussianization

Kernel-Based Nonlinear Anomaly Detection Via Union Dictionary for Hyperspectral Images

Improved Robust Kernel Subspace for Object-Based Registration and Change Detection

An Orthogonal Subspace Projection Anomaly Detection Algorithm Based on Adaptive Kernel Method

Iterative Graph Propagation for Hyperspectral Anomaly Detection

Spatial Filter Based Anomaly Detection Algorithm for Hyperspectral Imagery Kernel RX Detectors

Support Vector Data Description Based on Adaptive Anomaly Detection Method in Hyperspectral Imagery

Randomized kernels for large scale Earth observation applications

Real-Time Anomaly Detection Algorithm for Hyperspectral Remote Sensing by Using Recursive Polynomial Kernel Function

Anomaly detection and GPU implementation for spatial-spectral kernel spectral angle

An Efficient and Robust Framework for Hyperspectral Anomaly Detection

Using heterogeneous computing and edge computing to accelerate anomaly detection in remotely sensed multispectral images

Fast Real-Time Causal Linewise Progressive Hyperspectral Anomaly Detection Via Cholesky Decomposition

Feature Augmentation Improves Anomalous Change Detection for Human Activity Identification in Synthetic Aperture Radar Imagery

Approximate computing for onboard anomaly detection from hyperspectral images

A Nonparametric Estimator for Coherent Change Detection: The Permutational Change Detection

Improvement of Anomoly Detection Algorithms in Hyperspectral Images using Discrete Wavelet Transform