Doorway Detection Using Line Tracking and Vanishing Point Verification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing doorway detection methods for robots are limited in their ability to detect doorways from various angles and perspectives, particularly failing when top and bottom frames are not observed or when doors are unfamiliar, as they rely on shape and appearance information that is not generalizable to diverse door types and require extensive training data.

Innovation Solution

A system using line tracking and vanishing point-based verification, which extracts salient line segments from images, determines the existence of a vanishing point to switch between frontal and side doorway detection modules, and employs edge tracking and image intensity variance to validate door hypotheses, allowing for robust detection of doorways without the need for extensive offline training.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If training-based classification approaches are used for door detection, then detection accuracy improves for familiar door types, but the system fails to generalize to unseen door types and requires extensive training data

Engineering Contradiction:
Improvedoor detection accuracyVSAvoidgeneralization to unseen doors
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent changes the detection parameters from appearance-based features to geometry-based features (line orientations, vanishing points, relative positions). This allows the system to detect doors by their structural parameters rather than appearance, enabling generalization to unseen door types while maintaining detection accuracy

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent extracts only the essential geometric features (line segments, vanishing points, orientation angles) from door images, discarding appearance information. This extraction of critical geometric parameters enables the system to work with diverse door types without requiring training data

Inventive Principle:
Principle #2Taking out (Extraction)

2Reliability

If appearance and shape information are used for door validation, then detection performance improves on training data, but the learned features are not generalizable to diverse door types in real world

Engineering Contradiction:
Improvedoor detection reliabilityVSAvoidadaptability to diverse door types
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent validates doors by checking geometric parameters (line orientations within specific angle ranges, vanishing point distances, relative positions of door lines) rather than appearance features. This parameter-based validation ensures reliability across diverse door types without requiring learned features

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent creates a universal door detection framework that works for all door types by using common geometric properties (parallel lines, vanishing points, orientation angles) that are invariant to door appearance, material, or style variations

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

3Measurement precision

If manual parameter tuning is used for door validation, then detection can be optimized for specific scenarios, but parameters cannot be adjusted online based on distance to vanishing point or other information

Engineering Contradiction:
Improvedoor validation precisionVSAvoidonline parameter adjustment
Core Design Contradiction:
Measurement precisionVSEase of operation

Solution Approach 1:

The patent makes the validation parameters dynamic by adjusting them based on the distance to the vanishing point. As the robot moves closer or farther from doors, the acceptable ranges for line orientations and positions are automatically adjusted, eliminating the need for manual retuning while maintaining precision

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system uses feedback from vanishing point detection and line tracking to continuously adjust validation parameters. The detected geometric features feed back into the validation process, allowing automatic adaptation to different distances and viewing conditions

Inventive Principle:
Principle #23Feedback

4Measurement precision

If corner detection is required for frontal door detection, then door hypotheses can be validated, but the method cannot detect side doorways when top and bottom frames are not observed

Engineering Contradiction:
Improvedoor corner detection accuracyVSAvoiddetection of side doorways
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent extracts only the essential vertical line segments that define door positions, removing the requirement to detect corners or horizontal frames. This simplification allows detection of side doorways where top and bottom frames may be occluded or not visible

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent creates a universal detection approach that works for both frontal and side doorways by using only vertical line segments and vanishing point geometry, which are present in all door configurations regardless of viewing angle or frame visibility

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

Data Source

PatentUS10068336B1Generic frontal and side doorway detection with line tracking and vanishing point based verification
Publication Date: 2018.09.04 HRL LAB
  • US10068336B1 patent drawing
  • US10068336B1 patent drawing
  • US10068336B1 patent drawing

AI summary

Described is a system for frontal and side doorway detection. Salient line segments are extracted from an image frame captured of an indoor environment. Existence of a vanishing point in the image frame is determined. If a vanishing point is detected with a confidence score that meets or exceeds a predetermined confidence score, then the system performs side doorway detection via a side doorway detection module. If a vanishing point is detected with a confidence score below the predetermined confidence score, then the system performs frontal doorway detection via a frontal doorway detection module. A description of detected doorways is output and used by a mobile robot (Unmanned Aerial Vehicle) to autonomously navigate the indoor environment.