Vehicle Light-Spot Recognition for Unknown-Height Motion Discrimination

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing environment recognition devices face difficulties in distinguishing between stationary and moving light spots, especially when the height of light sources like street lights, traffic lights, and reflections is unknown, leading to erroneous detection and headlight control.

Innovation Solution

An environment recognition device that detects light spots at multiple points in time using an imaging device and determines their movement relative to the ground by calculating the variation in positions based on the host vehicle's movement, using a reference generation unit to estimate the moving amount of light spots and correlate it with the host vehicle's movement, without requiring the true height of the light spots.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the height of light spots is assumed to be on the road surface for tracking, then the tracking accuracy for vehicle lights is improved, but stationary objects like street lights and traffic lights are erroneously detected as vehicle lights

Engineering Contradiction:
Improvetracking accuracyVSAvoiddetection accuracy
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent changes the parameter of light spot height assumption from fixed road surface level to multiple possible heights. By calculating tracking results for various height assumptions and comparing them, the system can distinguish between stationary objects (which show inconsistent tracking across different height assumptions) and moving vehicle lights (which show consistent tracking patterns), thereby resolving the contradiction between tracking accuracy and detection reliability

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent introduces dynamic height adjustment in the tracking process. Instead of using a fixed height assumption, the system dynamically tests multiple height scenarios and selects the most appropriate one based on tracking consistency. This dynamic approach allows accurate tracking of vehicle lights at various heights while preventing false detection of stationary objects

Inventive Principle:
Principle #15Dynamics

2Productivity

If pattern shape matching is used to extract object candidates, then the detection speed is improved, but the ability to distinguish between vehicle lights and other light sources is reduced

Engineering Contradiction:
Improvedetection speedVSAvoidlight source discrimination accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent segments the detection process into two stages: first using pattern shape matching for rapid object candidate extraction, then applying height-based tracking analysis for precise light source discrimination. This segmentation maintains detection speed in the first stage while adding discrimination accuracy in the second stage, resolving the contradiction between productivity and measurement precision

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3438933B1Environment recognition device
Publication Date: 2023.10.04 ASTEMO LTD
  • EP3438933B1 patent drawingFigure 1
  • EP3438933B1 patent drawingFigure 2
  • EP3438933B1 patent drawingFigure 3

AI summary

This environment recognition device provides an environment recognition device with which it is possible to discriminate a stationary light spot and a moving light spot even when the height of the light spot concerned is unknown. This environment recognition device is characterized by being provided with: a light-spot detection unit (12) for detecting a light spot from each of images photographed at a plurality of points in time using a camera (C1) mounted onboard a host vehicle; and a moving-light-spot determination unit (15) for determining, on the basis of variation in the positions of light spots that are determined to be identical among the light spots detected from the images photographed at the plurality of points of time, whether the light spots are a light spot moving relative to the ground.