Multispectral Region Extraction for Narrowband Object Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing technologies are limited to detecting target objects that are present at specific positions in an imaging region, failing to effectively identify objects emitting light with a specific narrowband spectrum when they may be present at any position.

Innovation Solution

A region extraction device and method that identifies regions where target objects emitting light with a specific narrowband spectrum are not present, excluding these regions from the determination area, thereby improving detection performance by reducing false positives and processing time.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If detection is performed across the entire imaging region, then the possibility of detecting target objects at any position is improved, but processing time increases and false positives increase

Engineering Contradiction:
Improvedetection coverageVSAvoidprocessing time
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

The imaging region is segmented into multiple regions of interest (ROIs) based on spatial information from other wavelength bands. This segmentation allows the system to focus detection resources only on relevant regions rather than processing the entire imaging region, thereby reducing processing time while maintaining detection coverage.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system performs preliminary detection using light reception data from other wavelength bands before conducting the actual target object detection. This preliminary action identifies potential regions where target objects may be present, allowing the main detection process to be applied only to these pre-identified regions, thus reducing overall processing time.

Inventive Principle:
Principle #10Preliminary action

2Adaptability or versatility

If detection is performed across the entire imaging region, then the possibility of detecting target objects at any position is improved, but the number of false positives increases

Engineering Contradiction:
Improvedetection coverageVSAvoidfalse positive rate
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

By segmenting the imaging region into multiple ROIs based on spatial correlations across wavelength bands, the system reduces the area subject to false positives. Each ROI is independently evaluated, and false positives in non-ROI areas are eliminated since those areas are not subjected to target object detection.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system uses an intermediary determination region derived from other wavelength bands as a mediator to guide the target object detection process. This intermediary region acts as a filter that precedes the actual detection, reducing the probability of false positives by limiting detection to areas where target objects are likely to be present.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Productivity

If the determination region is reduced to specific known positions, then processing time is reduced, but the ability to detect objects at unknown positions is lost

Engineering Contradiction:
Improvedetection efficiencyVSAvoiddetection flexibility
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

The determination region is not fixed but dynamically determined based on the spatial distribution of light reception data across multiple wavelength bands. This dynamic adjustment allows the system to adapt to different scenarios and target object positions while maintaining efficient processing by focusing only on relevant regions.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system uses the same preliminary determination mechanism across all wavelength bands to universally identify potential target regions. This multi-functional approach allows the system to detect target objects regardless of their specific position or wavelength characteristics, enhancing detection flexibility while maintaining efficiency.

Inventive Principle:
Principle #6Universality (Multi-functionality)

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 enhances the detection of target objects by narrowing down the determination region, reducing erroneous detections and shortening processing time, while efficiently identifying objects like traffic signals using multispectral cameras.

Implementation Method 1

a target object that emits light having a light emission spectrum of a first narrowband

Methodology Applied
Scientific EffectLight emission spectrum detection: Absorption Spectroscopy

Implementation Method 2

images of a plurality of narrowbands including a second narrowband corresponding to the light emission spectrum of light emitted by the target object and a third narrowband different from the second narrowband

Methodology Applied
Scientific EffectNarrowband filtering: Filter (optical)

Data Source

PatentEP3961581B1Region extraction device, method thereof, object detection apparatus, and method thereof
Publication Date: 2026.04.15 FUJIFILM CORP
  • EP3961581B1 patent drawingFigure 1
  • EP3961581B1 patent drawingFigure 2
  • EP3961581B1 patent drawingFigure 3

AI summary

A region extraction device and a method thereof in a stage previous to detection of a target object, capable of appropriately extracting a region having a possibility of presence of a target object emitting light having a light emission spectrum of a specific narrowband as a determination region in an imaging region, and an object detection apparatus and a method thereof capable of efficiently detecting the target object using a region extraction result are provided. In a region extraction method, a plurality of images including an image of a second narrowband corresponding to a first narrowband of light emitted by a target object and an image of a third narrowband different from the second narrowband are acquired from a multispectral camera. Next, a region that emits light having a light emission spectrum other than a light emission spectrum of the first narrowband is determined as a non-determination region in an imaging region based on the plurality of acquired images, and one or a plurality of regions excluding the non-determination region from the imaging region are extracted as a determination region.