Articulated Machine Vision System Image Combination

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing vision systems for articulated machines face complexity in processing overlapping images from multiple cameras mounted on the front and rear sections, requiring additional control modules and complicating image processing when the front section moves relative to the rear section.

Innovation Solution

A vision system with a controller that includes image transformation modules to convert images from cameras on both sections into transformed images based on view type, and an image combination module to combine these images with reference to the articulation joint, eliminating the need for separate control modules to handle overlapping images.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Area of stationary object

If multiple cameras are mounted on the front and rear sections to monitor surroundings, then the monitoring coverage is improved, but image overlapping occurs and processing complexity increases

Engineering Contradiction:
Improvemonitoring coverageVSAvoidprocessing complexity
Core Design Contradiction:
Area of stationary objectVSDevice complexity

Solution Approach 1:

The image processing is segmented into distinct modules: a first image transformation module for front section cameras, a second image transformation module for rear section cameras, and an image combination module. Each module handles specific tasks independently, avoiding the need for complex overlapping detection and control across all images.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Image transformation is performed as a preliminary step before combination. The first and second transformed images are generated in advance through their respective transformation modules, converting raw camera images into a standardized format that eliminates overlapping issues before the combination stage.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If additional control modules are added to handle overlapping images, then image processing accuracy is improved, but device complexity increases

Engineering Contradiction:
Improveimage processing accuracyVSAvoidcontrol module quantity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The image combination module merges the first transformed image and the second transformed image into a single combined image that represents the complete surroundings of both front and rear sections. This unified approach achieves accurate monitoring without requiring separate control modules for each camera or overlapping management.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The image combination module serves multiple functions: it combines images from different sections, handles the articulation joint reference, and produces the final composite view. This multi-functional design eliminates the need for multiple specialized control modules while maintaining processing accuracy.

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

3Adaptability or versatility

If the front section moves relative to the rear section via articulation joint, then machine maneuverability is improved, but image alignment and combination difficulty increases

Engineering Contradiction:
Improvemachine maneuverabilityVSAvoidimage alignment complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system dynamically adapts to the articulation joint angle between front and rear sections. The image combination module uses the articulation joint angle information to correctly align and combine images from both sections, allowing the system to maintain accurate composite views regardless of the dynamic configuration of the articulated machine.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3048012B1Vision system and method of monitoring surroundings of machine
Publication Date: 2018.01.17 CATERPILLAR INC
  • EP3048012B1 patent drawingFigure 1
  • EP3048012B1 patent drawingFigure 2
  • EP3048012B1 patent drawingFigure 3

AI summary

A vision system (120) for a machine (100) is disclosed. The vision system (120) includes a first set of cameras (124) mounted on the first section (102) and a second set of cameras (128) mounted on the second section (104) of the machine (100). A controller (132) is coupled to the first set of cameras (124) and the second set of cameras (128). The controller (132) includes a first and a second image transformation modules (134a, 134b) to convert a first and a second images (FI, SI) captured by the first and the second set of cameras (124, 128) into a first and a second transformed image based at least on a view type, respectively. An image combination module (138) combines the first transformed image and the second transformed image to a combined image with reference to the articulation joint (110). The image combination module (138) communicates with a display device (122) to display the combined image. The combined image is indicative of surroundings of the first section (102) and the second section (104) of the machine (100).