Detecting Surface Changes on Venus: A GB2KI Thresholding Approach for Inter-mission SAR Images

The detection of surface changes on Venus is fundamental to understanding its ongoing volcanic activity and geological evolution. The upcoming EnVision and VERITAS missions, equipped with advanced SAR systems (VenSAR and VISAR), will image Venus more than 40 years after Magellan. This presents an unprecedented opportunity to detect active surface processes by comparing future high-resolution images with Magellan data. However, reliable inter-mission SAR change detection faces significant challenges due to differences in spatial resolution, wavelength, polarization, and viewing geometry.

A critical issue is the difference in spatial resolution between SAR images, which invalidates classical statistical models that assume identical equivalent number of looks (ENL) and leads to unreliable change detection results. In this work, we propose a novel change detection method by integrating the generalized beta prime (GB2) distribution into the Kittler-Illingworth (KI) minimum error thresholding framework, termed GB2KI. A modified criterion function is derived for optimal threshold selection. To address the class imbalance problem, we introduce an entropy-weighted maximum likelihood estimation method for robust parameter estimation. Additionally, a multiscale post-processing technique is developed to suppress noise patches and reduce false alarms in the final change detection map.

The proposed method is validated using both simulated and real SAR datasets. Simulations are conducted on Magellan Cycle 1 and Cycle 3 images by adding artificial changes with varying intensities, extents, and types to test the algorithm’s robustness. Further validation is performed using Earth observation data from the Holuhraun lava flow-field in Iceland. Two distinct datasets are analyzed, including Sentinel-1 images from different imaging modes (1-year interval) and Radarsat-1/Sentinel-1 images (15-year interval). Results demonstrate that our method achieves higher overall accuracy with significantly reduced false alarm rates compared to existing approaches.

This work provides a robust framework for inter-mission SAR change detection applicable to future Venus missions, enabling reliable identification of volcanic activity and other surface processes.