Work

A New Framework for the Detection and Removal of Glitches in InSight Seismic Data

Year: 2026

Type: article

Abstract: The InSight mission has collected a wealth of valuable seismic data from Mars. Analyses of these recordings have produced important new insights into the planet's interior structure and geological activity. However, the presence of numerous one‐sided transient signals, known as “glitches,” introduces difficulty and complexity into the processing of InSight seismic data. At present, no existing method is capable of perfectly detecting and removing all glitches, owing to their large number and varied origins. This study proposes a new glitch detection and removal approach by introducing a physically interpretable signal‐processing framework based on shock‐waveform decomposition (SWD) method. By exploiting the strong representational capability of the shock‐waveform, the proposed method decomposes raw seismic data into shock‐waveform components and identifies those associated with glitches and high‐frequency spikes for selective removal. Comparative analyses with established glitch detection and removal techniques demonstrate that the SWD‐based method achieves detection performance comparable to existing approaches, with intermediate detection sensitivity between that of the MPS method and those of the UCLA and GlitchNet methods and advantages of mathematical simplicity and a physical interpretation. In addition, the proposed approach effectively removes spikes occurring at glitch onset in the representative comparisons, resulting in cleaner processed signals in these cases.

Source: Earth and Space Science

DOI: 10.1029/2026ea005204

Authors: Andong Lu, Q. M. Li

Institution: University of Manchester

Language: English

Topic: Seismology and Earthquake Studies

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: gold

Estimated APC paid: $2,650