JUHE API Marketplace
DATASET
Open Source Community

HunanMultimodalDataset

This is a multimodal remote‑sensing dataset for Hunan Province in 2017, comprising Sentinel‑2, Sentinel‑1, and SRTM data. It contains 400 training images (256×256), and 50 validation and test images. The TRI in the training set is computed from SRTM via GDAL and can be used for knowledge reconstruction.

Updated 3/31/2024
github

Description

Dataset Overview

Dataset Name

HunanMultimodalDataset

Dataset Link

Dataset Link

Dataset Content

  • Year: 2017
  • Location: Hunan Province, China
  • Data Types: Multimodal remote‑sensing data, including Sentinel‑2, Sentinel‑1, and SRTM digital elevation data

Dataset Purpose

Used for land‑cover classification, especially with the proposed Domain Knowledge‑guided Deep Fusion Network (DKDFN).

Dataset Structure

  • Training Set: 400 images of 256×256, includes TRI (derived from SRTM via GDAL)
  • Validation & Test Sets: 50 images of 256×256 each

Data Processing

  • Label Pre‑processing: Provided code converts IGBP classes to 0‑6 categories (0: cultivated land, 1: forest, 2: grassland, 3: wetland, 4: water, 5: unused land, 6: built‑up area).

Dataset Features

  • Network Architecture: Multi‑head encoder with multi‑branch decoder supporting multitask learning (semantic segmentation and multimodal remote‑sensing index reconstruction).
  • Loss Function: Proposed Asymmetric Loss Function (ALF) optimized for minority classes.
  • Experimental Validation: Compared with six existing models (U‑Net, SegNet, PSPNet, DeepLab, HRNet, MP‑ResNet) demonstrating DKDFN’s superiority.

AI studio

Generate PPTs instantly with Nano Banana Pro.

Generate PPT Now

Access Dataset

Login to Access

Please login to view download links and access full dataset details.

Topics

Remote Sensing Technology
Geographic Information System (GIS)

Source

Organization: github

Created: 2/16/2022

Power Your Data Analysis with Premium AI Models

Supporting GPT-5, Claude-4, DeepSeek v3, Gemini and more.

Enjoy a free trial and save 20%+ compared to official pricing.