Agriculture-Vision Challenge 2024 -- The Runner-Up Solution for Agricultural Pattern Recognition via Class Balancing and Model Ensemble

TL;DR

This paper presents a solution for agricultural semantic segmentation in satellite images, addressing class imbalance with data augmentation, adaptive weighting, and post-processing, achieving second place in the CVPR 2024 challenge.

Contribution

It introduces a novel combination of class balancing techniques and model ensemble strategies specifically tailored for agricultural image segmentation.

Findings

Achieved a mean IoU of 0.547 on the challenge test set.

Effectively mitigated class imbalance in satellite image segmentation.

Secured second place in the Agriculture-Vision Challenge 2024.

Abstract

The Agriculture-Vision Challenge at CVPR 2024 aims at leveraging semantic segmentation models to produce pixel level semantic segmentation labels within regions of interest for multi-modality satellite images. It is one of the most famous and competitive challenges for global researchers to break the boundary between computer vision and agriculture sectors. However, there is a serious class imbalance problem in the agriculture-vision dataset, which hinders the semantic segmentation performance. To solve this problem, firstly, we propose a mosaic data augmentation with a rare class sampling strategy to enrich long-tail class samples. Secondly, we employ an adaptive class weight scheme to suppress the contribution of the common classes while increasing the ones of rare classes. Thirdly, we propose a probability post-process to increase the predicted value of the rare classes. Our methodology achieved a mean Intersection over Union (mIoU) score of 0.547 on the test set, securing second place in this challenge.