Multi-Camera Image Processing for Low-Bandwidth 3D Positioning

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image processing systems face challenges in estimating a three-dimensional position of an object when multiple cameras are connected to different calculation resources, as they require aggregating total image data, which can exceed transmission capacity, leading to estimation failures.

Innovation Solution

An image processing device that includes an arithmetic operation device to detect objects from multiple cameras, estimate distances, determine feature points, and collate feature amounts across images, using camera parameters to measure distances, thereby reducing data transmission requirements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If total image data from multiple cameras is aggregated to a single calculation resource, then three-dimensional position estimation can be performed, but data transmission capacity and storage capacity requirements increase significantly

Engineering Contradiction:
Improvethree-dimensional position estimation accuracyVSAvoiddata transmission quantity
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent segments the image processing task by dividing it into two stages: first, each calculation resource independently processes its local camera images to extract feature amounts; second, only these extracted feature amounts are transmitted and aggregated at a specified calculation resource. This segmentation reduces the data transmission quantity from transmitting entire images to transmitting only essential feature data, while still enabling three-dimensional position estimation.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent extracts only the necessary feature amounts from the complete image data at each calculation resource. Instead of transmitting all image pixels, the system identifies and transmits only the relevant feature quantities needed for three-dimensional position calculation. This extraction process significantly reduces data transmission requirements while preserving the essential information needed for accurate estimation.

Inventive Principle:
Principle #2Taking out (Extraction)

2Quantity of substance

If feature points are determined using estimated distance and camera parameters, then data transmission is reduced, but feature point allocation accuracy may be affected

Engineering Contradiction:
Improvedata transmission quantityVSAvoidfeature point allocation accuracy
Core Design Contradiction:
Quantity of substanceVSManufacturing precision

Solution Approach 1:

The patent performs preliminary distance estimation for each object using independently available information (object detection results and camera parameters) before feature point determination. This preliminary action allows each calculation resource to pre-calculate the necessary feature points based on the estimated distance, ensuring that only relevant feature amounts are extracted and transmitted. This approach maintains feature point allocation accuracy while reducing data transmission by eliminating unnecessary feature points.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4644832A1Image processing device
Publication Date: 2025.11.05 ASTEMO LTD
  • EP4644832A1 patent drawingFigure 1~2
  • EP4644832A1 patent drawingFigure 3
  • EP4644832A1 patent drawingFigure 4

AI summary

An image processing device includes an arithmetic operation device. The arithmetic operation device detects an object for each of a plurality of images imaged by a plurality of cameras, estimates a distance to the object for each of the plurality of images, determines a feature point that can be acquired in any of the plurality of images among feature points of the object using the estimated distance to the object for each of the plurality of images and camera parameters of the plurality of cameras, obtains a feature amount around the determined feature point for each of the plurality of images, collates feature amounts of the plurality of images, and measures the distance to the object based on a result of collating the feature amounts.