# Empirical Evaluation of UNet for Segmentation of Applicable Surfaces for Seismic Sensor Installation

## TL;DR

This paper evaluates the U-Net model for identifying suitable surfaces for seismic sensor installation using satellite imagery, providing a baseline for this new application.

## Contribution

The study introduces a novel dataset and evaluates U-Net with different encoders and band combinations for seismic sensor site selection.

## Key findings

- CSPDarknet53 encoder achieved the best performance with IoU = 0.534, Precision = 0.716, and Recall = 0.635.
- RGB + Near-Infrared bands at 10 m/pixel resolution provided the most robust results across configurations.
- Reducing the input stride improved segmentation of small linear objects like roads.

## Abstract

The deployment of wireless seismic nodal systems necessitates the efficient identification of optimal locations for sensor installation, considering factors such as ground stability and the absence of interference. Semantic segmentation of satellite imagery has advanced significantly, and its application to this specific task remains unexplored. This work presents a baseline empirical evaluation of the U-Net architecture for the semantic segmentation of surfaces applicable for seismic sensor installation. We utilize a novel dataset of Sentinel-2 multispectral images, specifically labeled for this purpose. The study investigates the impact of pretrained encoders (EfficientNetB2, Cross-Stage Partial Darknet53—CSPDarknet53, and Multi-Axis Vision Transformer—MAxViT), different combinations of Sentinel-2 spectral bands (Red, Green, Blue (RGB), RGB+Near Infrared (NIR), 10-bands with 10 and 20 m/pix spatial resolution, full 13-band), and a technique for improving small object segmentation by modifying the input convolutional layer stride. Experimental results demonstrate that the CSPDarknet53 encoder generally outperforms the others (IoU = 0.534, Precision = 0.716, Recall = 0.635). The combination of RGB and Near-Infrared bands (10 m/pixel resolution) yielded the most robust performance across most configurations. Reducing the input stride from 2 to 1 proved beneficial for segmenting small linear objects like roads. The findings establish a baseline for this novel task and provide practical insights for optimizing deep learning models in the context of automated seismic nodal network installation planning.