No abstract

1 , German Aerospace Center 2 , Planet Labs 3 Figure 1. Visualization of the DynamicEarthNet dataset. For a specific area of interest, we show two satellite observations, 2019-08-01 and 2019-08-31, as well as the corresponding monthly ground-truth annotation (top left). The complete dataset consists of daily samples in the range from 2018-01-01 to 2019-12-31. We consider 75 separate areas of interest, spread over six continents (top right).

Detecting changes on the ground in multitemporal Earth observation data is one of the key problems in remote sensing. In this paper, we introduce Sibling Regression for Optical Change detection (SiROC), an unsupervised method for change detection in optical satellite images with medium and high resolution. SiROC is a spatial context-based method that models a pixel as a linear combination of its distant neighbors.It uses this model to analyze differences in the pixel and its spatial context-based predictions in subsequent time periods for change detection. We combine this spatial context-based change detection with ensembling over mutually exclusive neighborhoods and transitioning from pixel to object-level changes with morphological operations. SiROC achieves competitive performance for change detection with medium-resolution Sentinel-2 and highresolution Planetscope imagery on four datasets. Besides accurate predictions without the need for training, SiROC also provides a well-calibrated uncertainty of its predictions.

No abstract

Figure 1: For a collection of satellite and street images, our method synthesizes the street view for each satellite input (right). It also simultaneously determines the geographic location of a query street image by matching it with the closest satellite image in the database (left→right). This is done in one single architecture which allows for end-to-end training.