Wheel Loader 3D Image Modeling for Sensorless Work Classification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Wheel loaders without sensors face challenges in accurately classifying work types due to disturbances during operations, making it difficult to detect the posture of work implements with high precision.

Innovation Solution

A system equipped with an imaging device and a computer that generates a three-dimensional model of the wheel loader's work implement, allowing for accurate classification of work types through image processing and machine learning models.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If image processing is used to classify work types without sensors, then device complexity is reduced, but measurement precision deteriorates due to disturbances during operations

Engineering Contradiction:
Improvesensor configurationVSAvoidwork type classification accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent introduces an intermediary coordinate transformation system that converts images from the camera coordinate system to the work implement coordinate system. This intermediary transformation process mediates between the raw image data and the final work type classification, enabling accurate measurement without requiring direct sensor mounting on the work implement. The coordinate transformation acts as a bridge that eliminates the need for complex sensor configurations while maintaining high classification accuracy.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces the mechanical sensor-based detection system with an optical image processing system. Instead of using physical sensors to directly detect work implement posture, the system uses camera images combined with coordinate transformation and machine learning algorithms to infer work types. This substitution reduces device complexity by eliminating physical sensors while compensating for measurement precision through computational methods.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If sensors are mounted on work implements to detect posture, then measurement precision is improved, but device complexity increases

Engineering Contradiction:
Improvework implement posture detectionVSAvoidsensor mounting and configuration
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent creates a virtual copy of the work implement posture information through image processing rather than using physical sensors. The system captures visual information via camera and reconstructs posture data through coordinate transformation and image analysis. This copying approach provides the necessary measurement precision while avoiding the complexity of mounting and configuring physical sensors on the work implement.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The coordinate transformation system serves as an intermediary that translates camera coordinate data into work implement coordinate data without requiring direct physical contact or sensor mounting. This intermediary layer enables precise posture detection by mathematically transforming the relationship between the camera view and the work implement orientation, thereby achieving sensor-level precision through computational means.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Measurement precision

If multiple imaging devices are used to reduce disturbances, then measurement precision is improved, but device complexity increases

Engineering Contradiction:
Improvework classification accuracyVSAvoidimaging device configuration
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent transitions from considering only the spatial position of the work implement to incorporating the camera's viewing angle and coordinate orientation as additional dimensions. By transforming images from the camera coordinate system to the work implement coordinate system, the method adds angular and orientational dimensions to the analysis. This dimensional transformation enables accurate work type classification using a single camera by fully utilizing the spatial and angular information contained in the images.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentEP3940154B1System including a work machine and a computer
Publication Date: 2024.08.21 KOMATSU LTD
  • EP3940154B1 patent drawingFigure 1
  • EP3940154B1 patent drawingFigure 2
  • EP3940154B1 patent drawingFigure 3(A)~3(F)

AI summary

An image clearly displaying a work machine is easily obtained. A posture data generation unit (163) estimates a posture of a work implement with respect to the body of the work machine in a captured image displaying the work machine. A motion state image generation unit (165) creates a three-dimensional model representing a stereoscopic shape of the work machine based on the posture of the work implement. A specific viewpoint image generation unit (166) creates image data including a two-dimensional image of the three-dimensional model, as viewed at a viewpoint position indicating a position of a viewpoint at which the three-dimensional model is virtually viewed.