Lens System for Random Dot Projection in Computer Vision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computer vision systems using two-dimensional camera inputs face challenges in accurately detecting objects at different distances due to regular dot patterns, leading to incorrect object positioning, as too many texture parts resemble the background, causing objects to disappear relative to the background.
Innovation Solution
A lens system comprising multiple light emitters arranged in a pattern, with a cluster of lenses that focus and project light in various directions, creating a random or semi-random light pattern detectable by cameras but not by humans, allowing for accurate object positioning determination using computer vision algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a regular dot pattern is used for illumination, then the lighting structure is simple, but objects may disappear relative to the background when at a particular distance, reducing measurement precision
Solution Approach 1:
The patent applies asymmetry by using a random or semi-random dot pattern instead of a regular grid pattern. This irregular arrangement ensures that the illumination pattern does not create periodic artifacts that could confuse the computer vision system, allowing objects at any distance to be distinguished from the background with high precision while maintaining relatively simple lighting hardware
2Measurement precision
If a random dot pattern is used for illumination, then object position detection precision is improved, but the lighting structure becomes more complex
Solution Approach 1:
The patent segments the illumination into multiple discrete light sources arranged in a random or semi-random pattern. Each light source contributes to the overall illumination, and the segmented arrangement creates a unique texture pattern that enhances object detection precision without requiring a monolithic complex lighting structure
Solution Approach 2:
The patent changes the spatial distribution parameter of the light sources from a regular grid to a random or semi-random arrangement. This parameter change transforms the illumination pattern to provide better depth cues for computer vision while keeping the physical lighting structure relatively simple and manageable
3Adaptability or versatility
If invisible infrared light is used, then the illumination pattern is detectable by cameras but not by human observers, but the light source complexity increases
Solution Approach 1:
The patent substitutes visible light with infrared light, changing the electromagnetic spectrum band used for illumination. This substitution allows the system to operate in an invisible spectrum that cameras can detect, enhancing adaptability for computer vision applications while using standard infrared LED technology that does not significantly increase light source complexity
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
The system enables precise object tracking and positioning in space by generating unique illumination patterns that improve the performance of stereo algorithms, ensuring all objects have sufficient texture for accurate depth determination, even in conditions with high ambient illumination.
Implementation Method 1
a cluster of lenses focuses and projects light from the emitters in numerous directions
Implementation Method 2
The emitters are arranged in a pattern. The system further includes a cluster of lenses located in front of the light source
Implementation Method 3
The system may alternatively (or additionally) include a condenser lens located between the light source and the cluster of lenses. The condenser lens concentrates light from the light source to the cluster of lens
Implementation Method 4
the system may include a cluster of infrared light emitting diodes (LEDs). The light emitting diodes generate infrared light that is detectable by the camera but not by the eye of a human observer
Data Source
AI summary
A fragmented lens system for creating an invisible light pattern useful to computer vision systems is disclosed. Random or semi-random dot patterns generated by the present system allow a computer to uniquely identify each patch of a pattern projected by a corresponding illuminator or light source. The computer may determine the position and distance of an object by identifying the illumination pattern on the object.