Urban Flood Inundation Mapping with Cloud Removal and Super-Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing flood disaster monitoring methods using polar-orbiting satellites and radar face challenges in achieving timely and comprehensive urban flood inundation monitoring due to cloud interference, limited spatial resolution, and long revisit cycles, which hinder accurate and real-time identification of flood extents.
Innovation Solution
A method combining Fengyun-4 satellite optical imagery with Sentinel-2 radar imagery using homomorphic filtering and super-resolution convolutional neural networks to remove clouds and enhance spatial resolution, enabling accurate extraction of urban flood inundation extents.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If polar-orbiting satellite optical images are used for flood monitoring, then spatial resolution is improved, but cloud cover interference increases making surface information acquisition difficult
Solution Approach 1:
The patent combines polar-orbiting satellite optical images with geosynchronous satellite infrared images to create a composite monitoring system. The optical images provide high spatial resolution for detailed flood extent mapping, while the infrared images penetrate cloud cover to provide continuous monitoring capability, thereby resolving the contradiction between spatial resolution and cloud interference.
Solution Approach 2:
The patent introduces infrared radiation as an intermediary mechanism to overcome cloud cover. Since infrared radiation can penetrate thin cloud layers, it serves as a mediator that allows surface information acquisition even when optical bands are blocked by clouds, thus maintaining monitoring capability under cloudy conditions.
2Area of stationary object
If polar-orbiting satellites are used for flood monitoring, then spatial coverage is improved, but revisit period increases causing missed flood events
Solution Approach 1:
The patent merges the observation capabilities of polar-orbiting satellites (wide coverage) with geosynchronous satellites (frequent revisits). This combination allows the system to maintain comprehensive spatial coverage while reducing the effective revisit period, ensuring timely detection of flood events without missing occurrences.
Solution Approach 2:
The patent utilizes the periodic orbital characteristics of both satellite types to create a multi-frequency monitoring schedule. Polar-orbiting satellites provide periodic wide-area coverage, while geosynchronous satellites provide continuous or near-continuous monitoring of specific regions, together creating a dense temporal sampling pattern that captures flood events.
3Reliability
If SAR is used for flood monitoring, then all-weather observation capability is improved, but scanning range decreases limiting monitoring coverage
Solution Approach 1:
The patent combines SAR's all-weather radar imaging capability with optical satellite imaging capability. The SAR provides reliable data under cloudy and rainy conditions, while optical satellites provide high-resolution data during clear periods, together creating a comprehensive all-weather monitoring system with extended effective coverage.
Solution Approach 2:
The patent creates a multi-functional monitoring system that can operate under various weather conditions using different satellite types. The system universally monitors flood events regardless of weather by switching between or combining radar and optical data sources, thereby achieving both all-weather capability and comprehensive coverage.
Data Source
AI summary
A method for extracting urban flood inundation extent based on remote sensing images includes: acquiring and preprocessing a Fengyun-4 satellite remote sensing image to obtain a preprocessed remote sensing image, and dividing the preprocessed remote sensing image into a thin cloud area and a thick cloud area; removing thin clouds and thick clouds in the preprocessed remote sensing image; extracting a first water body area from the cloud-removed remote sensing image to obtain a first urban inundation extent; acquiring a radar image at a same region and a same time as the remote sensing image, and extracting a second water body area from the radar image to obtain a second urban inundation extent; and training a super-resolution convolutional neural network according to the first urban inundation extent and the second urban inundation extent, and using a trained super-resolution convolutional neural network to obtain a reconstructed higher-resolution urban flood inundation extent.