Water Surface Image Compositing for Reflection Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image processing technologies struggle to effectively reduce the effect of reflected light on water surfaces, especially when the angle between the optical axis and the water surface normal is large, limiting the ability to capture accurate images of water quality and quality distribution.

Innovation Solution

An image processing apparatus and method that uses a combination of optical filters and image processing techniques to capture images at different wavelength ranges, including 670 nm and 700 nm, and artificially generates waves to reduce the effect of reflected light, allowing for accurate water quality distribution imaging.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If multiple images are captured at different positions and depths of field, then subject isolation and background separation are improved, but the number of images increases and processing complexity increases

Engineering Contradiction:
Improvesubject isolation precisionVSAvoidimage processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the image processing task into distinct modules: depth map generation module, segmentation module, and synthesis module. Each module processes specific aspects of the image independently, allowing complex subject isolation to be achieved through coordinated simple operations rather than a single complex algorithm.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces depth maps as an intermediary representation between the captured images and the final segmented result. Instead of directly segmenting complex multi-depth images, the system uses depth information as a mediator to guide the segmentation process, simplifying the overall computation while maintaining precision.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If images are captured with different depths of field, then subject isolation is improved, but alignment accuracy decreases due to perspective changes

Engineering Contradiction:
Improvesubject isolation precisionVSAvoidimage alignment accuracy
Core Design Contradiction:
Measurement precisionVSManufacturing precision

Solution Approach 1:

The patent changes the parameter space from direct pixel-based alignment to depth-map-based alignment. By transforming images into depth maps and back, the system adjusts the representation parameters to accommodate perspective changes while maintaining alignment accuracy across different depth of field settings.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent performs preliminary depth map generation and feature extraction before the actual alignment and synthesis operations. This preliminary processing prepares the data in advance, allowing subsequent alignment operations to proceed more accurately and efficiently despite perspective variations.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If multiple images are captured and processed, then subject isolation is improved, but processing time increases

Engineering Contradiction:
Improvesubject isolation precisionVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent implements continuous processing where depth maps are generated, features are extracted, and images are synthesized in an overlapping manner rather than sequentially completing all operations. This continuous action reduces total processing time while maintaining the precision benefits of multi-image processing.

Inventive Principle:
Principle #20Continuity of useful action

Solution Approach 2:

The patent applies partial processing to regions where depth information is most critical, rather than uniformly processing all image areas. This selective approach reduces overall processing time while maintaining subject isolation precision in the most important regions.

Inventive Principle:
Principle #16Partial or excessive action

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

The solution achieves a significant reduction in the effect of reflected light across a wide range, enabling accurate water quality distribution imaging and chlorophyll a concentration mapping, even when natural waves are absent, by employing optical filters and image processing algorithms.

Implementation Method 1

a first optical filter allowing transmission of light having a first wavelength range, a second optical filter allowing transmission of light having a second wavelength range different from the first wavelength range

Methodology Applied
Scientific EffectOptical filtering by wavelength: Filter (optical)

Implementation Method 2

the luminous flux incident through the first optical filter, the second optical filter, and the third optical filter is selectively received by pupil separation

Methodology Applied
Scientific EffectPupil separation:

Implementation Method 3

a polarization filter through which a luminous flux of the same scene including the water surface having the wave passes

Methodology Applied
Scientific EffectPolarization filtering: Polarisation

Data Source

PatentEP3618419B1Image processing device, image processing method, and program
Publication Date: 2023.10.04 FUJIFILM CORP
  • EP3618419B1 patent drawingFigure 1
  • EP3618419B1 patent drawingFigure 2
  • EP3618419B1 patent drawingFigure 3

AI summary

An object of the present invention is to provide an image processing apparatus, an imaging apparatus, an image processing method, and a program capable of generating an image in which an effect of reflected light on a water surface is reduced in a wide range. An image processing apparatus 500 includes an image obtaining unit 501 that obtains a plurality of frame images of the same scene including a water surface having a wave, a brightness value calculation unit 503 that calculates a brightness value corresponding to a minute region constituting each frame image based on the plurality of frame images, a pixel value extraction unit 505 that extracts a pixel value corresponding to the minute region of a frame image in which the brightness value of the minute region is a brightness value lower than the brightness value of the minute region of another frame image in the plurality of frame images, a storage unit 507 that stores the pixel value extracted by the pixel value extraction unit 505, and a composite image generation unit 509 that generates a composite image corresponding to the scene based on the pixel value stored in the storage unit 507.