CloudMatch: Weak-to-Strong Consistency Learning for Semi-Supervised Cloud Detection

TL;DR

CloudMatch introduces a semi-supervised learning framework for cloud detection that leverages view-consistency and scene-mixing augmentations to effectively utilize unlabeled remote sensing imagery, capturing cloud pattern diversity.

Contribution

The paper presents CloudMatch, a novel semi-supervised approach that combines view-consistency learning with scene-mixing augmentations to improve cloud detection accuracy.

Findings

Achieves high performance in semi-supervised cloud detection tasks.

Effectively captures structural diversity and contextual variability of clouds.

Utilizes unlabeled data efficiently to enhance model generalization.

Abstract

Due to the high cost of annotating accurate pixel-level labels, semi-supervised learning has emerged as a promising approach for cloud detection. In this paper, we propose CloudMatch, a semi-supervised framework that effectively leverages unlabeled remote sensing imagery through view-consistency learning combined with scene-mixing augmentations. An observation behind CloudMatch is that cloud patterns exhibit structural diversity and contextual variability across different scenes and within the same scene category. Our key insight is that enforcing prediction consistency across diversely augmented views, incorporating both inter-scene and intra-scene mixing, enables the model to capture the structural diversity and contextual richness of cloud patterns. Specifically, CloudMatch generates one weakly augmented view along with two complementary strongly augmented views for each unlabeled image: one integrates inter-scene patches to simulate contextual variety, while the other employs intra-scene mixing to preserve semantic coherence. This approach guides pseudolabel generation and enhances generalization. Extensive experiments show that CloudMatch achieves good performance, demonstrating its capability to utilize unlabeled data efficiently and advance semi-supervised cloud detection.