Event Sensor Noise Removal for Structured Light Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High illumination conditions, such as those under the sun, lead to flare defects in structured light systems, causing noise in the object recognition system, which burdens subsequent signal processing.
Innovation Solution
An object recognition system that includes a light source section projecting dot light with a predetermined pattern, an event detection sensor detecting changes in pixel luminance, and a signal processor that identifies and removes noise events by processing multiple successive pixels as noise, with adjustable thresholds and directions for noise detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If event detection sensor is used to detect dot light pattern under high illumination, then object recognition capability is improved, but noise from flare increases
Solution Approach 1:
The patent segments the pixel array into groups of successive pixels and evaluates noise based on the collective behavior of each group rather than individual pixels. By dividing the pixel array into multiple groups and analyzing event occurrence patterns within each group, the system can distinguish between random noise and genuine dot light pattern events, thereby removing flare noise while preserving object recognition accuracy.
Solution Approach 2:
The patent extracts and removes noise events from the detected signal by identifying patterns characteristic of flare (multiple successive pixels detecting events within a short time period) and eliminating these identified noise components from the final object recognition data, thereby separating the harmful flare noise from the useful dot light pattern information.
2Measurement precision
If noise removal processing is applied to remove flare events, then signal processing accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent applies different processing treatments to different regions of the pixel array by evaluating each group of successive pixels independently. Instead of applying a uniform complex filtering algorithm across the entire array, the system uses local evaluation of event patterns within each pixel group, determining whether to remove events based on local noise characteristics, thereby reducing overall processing complexity while maintaining accuracy.
Solution Approach 2:
The patent performs preliminary noise identification and removal before final object recognition processing. By pre-identifying and removing flare noise events based on simple temporal and spatial patterns (multiple successive pixels detecting events within a threshold period), the system reduces the processing burden on subsequent recognition algorithms, thereby improving overall efficiency without requiring complex processing in later stages.
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
Effectively reduces noise from flare under high illumination, enhancing the accuracy of event information derived from the dot pattern light, thereby reducing the processing burden and improving system performance.
Implementation Method 1
a light source section that irradiates a subject with dot light having a predetermined pattern
Implementation Method 2
an event detection sensor that receives the dot light having the predetermined pattern reflected by the subject and detects, as an event, that a change in luminance of a pixel exceeds a predetermined threshold
Data Source
AI summary
An object recognition system of the present disclosure includes: a light source section that irradiates a subject with dot light having a predetermined pattern; an event detection sensor that receives the dot light having the predetermined pattern reflected by the subject and detects, as an event, that a change in luminance of a pixel exceeds a predetermined threshold; and a signal processor that performs, in a case where a plurality of successive pixels in a pixel array section of the event detection sensor detects occurrence of an event in a certain period, processing of removing the event as noise, the plurality of successive pixels being equal to or greater than a predetermined number of pixels.