Underwater Object Detection via Wavelet Transform Subtraction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Images acquired through 3-chip cameras over water surfaces are hindered by noise and glints, making it difficult to detect objects below the water's surface due to the presence of sensor noise and electronic circuit noise.

Innovation Solution

The method involves acquiring image data in separate spectral regions, converting it into intensity value arrays, transforming these arrays into two-dimensional discrete wavelet transform arrays, and then subtracting them to reduce noise and glints, thereby enhancing the detection of objects below the water's surface.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If image data is acquired through 3-chip cameras over water surfaces, then image data is obtained, but noise and glints make it difficult to detect objects below the water's surface

Engineering Contradiction:

Improvedetection capabilityVSAvoidnoise and glints

Core Design Contradiction:

Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The image processing is segmented into multiple spectral regions (e.g., blue, green, red channels) that are processed separately. Each spectral region is transformed into wavelet domains independently, allowing targeted noise and glint removal in specific frequency ranges before combining the processed channels to enhance object detection

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Wavelet transform serves as an intermediary mathematical tool that transforms image data from spatial domain to frequency domain. This intermediary transformation enables the separation and removal of noise and glint components while preserving the underlying object information, which is then transformed back to produce enhanced images

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If image processing is performed to minimize noise and glints, then detection capability is improved, but processing complexity increases

Engineering Contradiction:

Improvedetection capabilityVSAvoidprocessing complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The wavelet transform process extracts and isolates specific frequency components (noise and glints) from the image data. By taking out these unwanted components separately in the frequency domain and removing them, the processing achieves noise reduction without requiring complex spatial filtering operations

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

Traditional spatial domain noise filtering methods are replaced with wavelet transform-based frequency domain processing. This substitution provides a more efficient and systematic approach to noise and glint removal, reducing processing complexity while improving detection capability

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS7869620B2Methods for processing images to detect objects below the surface of the water

Publication Date: 2011.01.11 MICRO BIZ VENTURES LLC

AI summary

The present invention is a processing method to increase the detection of objects below the surface of the water comprising the steps of: acquiring image data in separate spectral regions simultaneously and converting the image data into intensity value arrays for each spectral region, transforming the intensity value arrays into two-dimensional discrete wavelet transform arrays for each spectral region, and subtracting in a pair wise manner the two-dimensional discrete wavelet transform arrays thereby obtaining images showing an object below the surface of the water.