High-Resolution Multispectral Dataset for Semantic Segmentation

TL;DR

This paper introduces a high-resolution multispectral dataset with labels to facilitate the development and benchmarking of semantic segmentation algorithms for non-RGB remote sensing imagery, addressing a key data gap.

Contribution

The paper provides a publicly available, pre-split high-resolution multispectral dataset with labels, enabling standardized evaluation of semantic segmentation methods for non-RGB data.

Findings

Dataset enables benchmarking of segmentation algorithms.

Standardized training/testing splits facilitate fair comparisons.

Supports development of algorithms for multispectral imagery.

Abstract

Unmanned aircraft have decreased the cost required to collect remote sensing imagery, which has enabled researchers to collect high-spatial resolution data from multiple sensor modalities more frequently and easily. The increase in data will push the need for semantic segmentation frameworks that are able to classify non-RGB imagery, but this type of algorithmic development requires an increase in publicly available benchmark datasets with class labels. In this paper, we introduce a high-resolution multispectral dataset with image labels. This new benchmark dataset has been pre-split into training/testing folds in order to standardize evaluation and continue to push state-of-the-art classification frameworks for non-RGB imagery.