Vehicle Object Detection Excluding Bar-Shaped Obstacles

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Collision avoidance systems in vehicles often incorrectly detect bar-shaped objects like those at ETC gates or railroad crossing gates, leading to unnecessary alarms and driver discomfort, as they fail to accurately differentiate these obstacles from other objects.

Innovation Solution

An information processing apparatus with an object region detecting unit, local region detecting unit, and object specifying unit is used to accurately identify and exclude bar-shaped objects by analyzing distance and luminance information, employing disparity images and feature amounts to distinguish these objects from other detection targets.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the collision avoidance system detects all objects in the detection range, then the detection coverage is improved, but false detection of bar-shaped objects (ETC gates, railroad crossing gates) causes unnecessary alarms and driver discomfort

Engineering Contradiction:
Improvedetection accuracyVSAvoidfalse alarms
Core Design Contradiction:
ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The object detection process is segmented into multiple stages: initial object region detection, bar-shaped object candidate identification, and final verification. By dividing the detection process into discrete segments with specific criteria for each stage, the system can identify and exclude bar-shaped objects while maintaining comprehensive detection of other objects.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces height information as an additional dimension for object analysis. By examining the vertical extent and position of detected objects, the system can distinguish bar-shaped objects (which have characteristic height profiles) from other objects, enabling accurate exclusion of false detection targets.

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

2Object-generated harmful factors

If the system excludes bar-shaped objects from detection targets, then false alarms are reduced, but the ability to detect actual obstacles may be compromised

Engineering Contradiction:
Improvefalse alarmsVSAvoiddetection accuracy
Core Design Contradiction:
Object-generated harmful factorsVSReliability

Solution Approach 1:

The system performs preliminary identification of bar-shaped objects using specific geometric criteria (aspect ratio, height, position) before final object detection. This preliminary action allows the system to pre-filter out false detection targets while maintaining sensitivity to actual obstacles that do not match the bar-shaped profile.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent replaces simple geometric exclusion with a more sophisticated detection mechanism that uses multiple features (height, width, position, aspect ratio) to identify bar-shaped objects. This substitution enables more accurate discrimination between actual obstacles and false detection targets.

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

Data Source

PatentEP3545465B1Information processing apparatus, imaging apparatus, apparatus control system, movable object, information processing method, and information processing program
Publication Date: 2024.04.10 RICOH CO LTD
  • EP3545465B1 patent drawingFigure 1~2
  • EP3545465B1 patent drawingFigure 3~5
  • EP3545465B1 patent drawingFigure 6~7

AI summary

According to an aspect of the present invention, an information processing apparatus includes an object region detecting unit, a local region detecting unit, an object specifying unit. The object region detecting unit is configured to detect an object region based on one of distance information and luminance information. The local region detecting unit is configured to, when a divided area obtained by dividing the detected object region meets a predetermined condition, detect the divided area as a local region. The object specifying unit is configured to specify, as a specification target object, the object region in which the local region is continuously detected.