Dot-Pattern Object Recognition With Event Noise Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object recognition systems using structured light methods with event detection sensors produce noise events along with true events, increasing processing loads.
Innovation Solution
An object recognition system that includes a light source emitting a predetermined pattern, an event detection sensor to detect luminance changes, and a signal processor to remove noise events from detected information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If an event detection sensor is used to detect dot light patterns, then detection speed and sensitivity are improved, but noise events are generated increasing processing load
Solution Approach 1:
The patent applies preliminary action by performing noise event removal immediately after event detection, before subsequent signal processing steps. The signal processor is configured to remove noise events based on temporal and spatial correlations, preparing clean data in advance for downstream processing. This preliminary filtering reduces the burden on subsequent processing stages while maintaining the high detection speed capability of the event detection sensor.
2Loss of information
If all detected event information is processed, then no information is lost, but processing time and computational resources increase
Solution Approach 1:
The patent applies the extraction principle by selectively removing noise events from the detected event information based on specific criteria. The signal processor identifies and extracts true events from the mixed signal by analyzing temporal correlations (events occurring at expected intervals) and spatial correlations (events following the dot light pattern geometry). This extraction process removes unnecessary noise events while preserving all meaningful information, significantly reducing processing time without sacrificing information completeness.
Solution Approach 2:
The patent applies local quality by applying different processing strategies to different types of events based on their characteristics. True events originating from dot light patterns are processed with full detail, while noise events are identified and removed through correlation analysis. The signal processor adapts its processing intensity locally - performing comprehensive analysis for potential true events and rapid rejection for noise events - thereby optimizing the balance between information retention and processing efficiency.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Reduces processing loads by effectively filtering out noise events, enhancing the efficiency of subsequent signal processing.
Implementation Method 1
a light source emitting dot light having a predetermined pattern to an object
Implementation Method 2
an event detection sensor receiving the dot light having the predetermined pattern reflected from the object and detecting a fact that a change in luminance of a pixel has exceeded a predetermined threshold as an event
Data Source
AI summary
An object recognition system of the disclosure includes a light source emitting dot light having a predetermined pattern to an object, an event detection sensor receiving the dot light having the predetermined pattern reflected by the object and detecting the fact that a change in luminance of a pixel has exceeded a predetermined threshold as an event, and a signal processor removing information other than event information originated from the dot light emitted from the light source and having the predetermined pattern among event information detected by the event detection sensor.