Object Detection Over Water Using NDVI Spectral Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current object detection systems over water face challenges such as high false detection rates due to atmospheric conditions and time-consuming processing, which can hinder search and rescue or surveillance missions.
Innovation Solution
The system uses normalized difference vegetation index (NDVI) by acquiring image data from a storage system, determining a metric for each pixel based on radiance in the NIR and red bands, and detecting objects when the metric satisfies a criteria, thereby reducing processing time and false alarms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional object detection methods are used over water, then detection accuracy can be maintained, but processing time becomes excessive and false detection rate increases
Solution Approach 1:
The patent transforms the detection approach by changing the parameter space from standard grayscale or color image analysis to spectral radiance analysis across multiple bands (visible, near-infrared, thermal infrared). By computing NDVI and other spectral indices, the system identifies objects based on their unique spectral signatures, which remain stable under varying atmospheric conditions. This parameter transformation enables rapid processing while maintaining high detection reliability and reducing false positives.
Solution Approach 2:
The patent introduces spectral indices (NDVI, NDWI, NDBI) as intermediary metrics that mediate between raw multi-band radiance data and final object detection decisions. These indices serve as robust intermediaries that are less sensitive to atmospheric variations compared to direct pixel intensity comparisons, thereby reducing false detections while enabling faster processing through simplified threshold-based classification.
2Measurement precision
If user intervention or extensive model training is applied, then detection accuracy improves, but processing time increases significantly
Solution Approach 1:
The patent implements a self-service detection system that automatically processes multi-band satellite imagery through predefined spectral index calculations and threshold-based classification algorithms. The system does not require manual user intervention or time-consuming model training phases. Instead, it autonomously computes NDVI, NDWI, and NDBI indices, compares them against calibrated thresholds, and generates detection results directly, achieving both high accuracy and rapid processing.
Solution Approach 2:
The patent replaces complex mechanical or manual detection processes (user intervention, iterative model training) with automated computational physics-based methods. By substituting human-in-the-loop processing with algorithmic spectral analysis, the system eliminates time-consuming manual steps while maintaining or improving detection accuracy through objective, reproducible spectral thresholding.
3Productivity
If standard image processing techniques are used, then processing is faster, but false detections increase due to atmospheric conditions
Solution Approach 1:
The patent moves the detection problem from two-dimensional spatial image analysis to five-dimensional spectral space by utilizing radiance values across multiple wavelength bands (visible red, near-infrared, thermal infrared). This dimensional expansion provides additional discriminatory information that enables the system to distinguish true objects from atmospheric artifacts, reducing false positives while maintaining fast processing through efficient spectral index computations.
Solution Approach 2:
The patent changes the detection parameters from standard spatial intensity metrics to spectral radiance ratios and differences (NDVI, NDWI, NDBI). These transformed parameters are inherently more robust to atmospheric conditions because they normalize out common atmospheric effects. The system processes these transformed parameters rapidly using simple arithmetic operations and threshold comparisons, achieving both speed and reliability.
Data Source
AI summary
A method and a system for object detection over water are provided. The method includes acquiring, from a storage system, image data associated with a target area. The image data includes radiance of a plurality of pixels in a first spectral band and in a second spectral band. The method also includes determining a metric corresponding to a pixel of the plurality of pixels as a function of the radiance in the first spectral band and the radiance in the second spectral band and detecting an object in the target area in response to a determination that the metric satisfies a criteria.