The monitoring of semi-natural ecosystems has become increasingly critical due to the rising impact of ecological disturbances, including natural disasters and unauthorized human-made constructions.

On this basis, we propose the denoising autoencoder based on the broad learning system (DBLS-AE), which sufficiently learns the anomalous patterns, achieving efficient anomaly detection with low ...