Working Machine Shadow Profile Extraction for Obstacle Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The complexity of a working machine's shadow on uneven terrain poses challenges for obstacle detection systems, leading to false recognition and decreased efficiency, as existing methods require extensive data storage and processing to distinguish the machine's shadow from obstacles, especially in off-road environments where shadows dynamically change.

Innovation Solution

A surroundings monitoring system that uses a monocular camera and image processing units to extract characteristic patterns, excluding the shadow profile from the obstacle detection target, allowing for accurate detection of objects within the shadow without relying on pre-stored shadow boundary patterns or color analysis.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a plurality of shadow boundary patterns are prepared in advance and stored in a memory apparatus, then shadow recognition accuracy may be improved, but device complexity and initial cost increase significantly

Engineering Contradiction:
Improveshadow recognition accuracyVSAvoiddata storage requirement
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts characteristic amounts (features) from the working machine image in advance and stores only these extracted features along with their corresponding ground regions, rather than storing complete shadow boundary patterns. This preliminary extraction action reduces the data storage requirement while maintaining recognition accuracy.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent extracts only the essential characteristic amounts (features) from the complex shadow boundary patterns and stores these extracted features. This extraction process removes unnecessary data while preserving the key information needed for accurate shadow recognition, thereby reducing device complexity.

Inventive Principle:
Principle #2Taking out (Extraction)

2Measurement precision

If countless shadow boundary patterns are prepared for every profile and road surface shape, then recognition accuracy may be improved, but loss of time and processing complexity increase

Engineering Contradiction:
Improveown vehicle shadow recognition accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent changes the representation parameters from complete shadow boundary patterns to extracted characteristic amounts (features). This parameter transformation reduces the data dimensionality and enables faster processing while maintaining recognition accuracy, thus reducing loss of time.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

Instead of storing and processing numerous complete shadow boundary patterns, the patent creates a simplified copy representation using only the essential characteristic amounts extracted from the images. This copying approach reduces processing complexity and time while preserving the necessary recognition information.

Inventive Principle:
Principle #26Copying

3Area of stationary object

If the camera is installed at a high position with great depression angle to monitor surroundings of the huge working machine, then monitoring coverage is improved, but shadow frequency and detection difficulty increase

Engineering Contradiction:
Improvemonitoring coverage areaVSAvoidobject detection difficulty in shadow
Core Design Contradiction:
Area of stationary objectVSDifficulty of detecting and measuring

Solution Approach 1:

The patent replaces the traditional mechanical approach of avoiding shadows by adjusting camera position with an image processing approach. By extracting characteristic amounts and comparing them with stored features, the system can accurately detect objects even when they appear in shadows, thus substituting mechanical constraints with computational solutions.

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

Data Source

PatentEP3136341B1Surroundings monitoring system for working machine
Publication Date: 2020.03.04 HITACHI CONSTRUCTION MACHINERY CO LTD
  • EP3136341B1 patent drawingFigure 1
  • EP3136341B1 patent drawingFigure 2
  • EP3136341B1 patent drawingFigure 3

AI summary

A surroundings monitoring system for a working machine is provided by which, even if an image obtained by imaging surroundings of the working machine includes an own vehicle shadow of a complicated shape, it can be prevented that the presence of the own vehicle shadow has an influence on detection of an object existing around the working machine. The surroundings monitoring system for a working machine includes: a monocular camera 6 that picks up an image of the surroundings of the working machine; a characteristic pattern extraction unit 170 that extracts characteristic patterns in the image based on a characteristic amount of the image; a shadow profile extraction unit 40 that extracts a profile of a region, which can be regarded as a shadow of the working machine in the image, based on the characteristic amount of the image; and an object detection unit 180 that detects an obstacle existing around the working machine based on the remaining characteristic patterns obtained by excluding a shadow profile characteristic patterns positioned on the profile extracted by the shadow profile extraction unit from the characteristic patterns extracted by the characteristic pattern extraction unit.