Structured Light Touch Recognition Using Dual-Pitch Epipolar Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing image recognition technology for detecting finger contact with a screen has low position detection accuracy in the depth direction, resulting in low touch recognition accuracy.
Innovation Solution
An image recognition device that displays alternating patterns with different pitches on an epipolar line, using a combination of a first pattern with a first pitch and a second pattern with a second pitch, shorter than twice the first pitch, to enhance position detection and distinguish phase wrapping, thereby improving touch recognition accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single pitch pattern is used for structured light projection, then the device complexity is low, but the position detection accuracy in the depth direction is insufficient
Solution Approach 1:
The patent divides the structured light pattern into multiple pitch variations (first pitch and second pitch) projected at different time points. This segmentation of the pattern into distinct pitch components enables accurate depth position detection by comparing phase differences, while maintaining manageable device complexity through temporal multiplexing of the patterns.
Solution Approach 2:
The patent employs periodic projection of alternating first and second patterns with different pitches at different time points. This periodic action allows the system to capture multiple phase information sets over time, improving depth detection accuracy through temporal comparison while keeping the projection device operationally simple through rhythmic pattern switching.
2Measurement precision
If the second pitch is made very short to improve detection accuracy, then the position detection accuracy improves, but phase wrapping occurs more frequently
Solution Approach 1:
The patent changes the pitch parameter of the structured light pattern by using two different pitch values (first pitch and second pitch) alternately. This parameter variation allows the system to detect depth positions accurately while the specific constraint that the second pitch is shorter than twice the first pitch optimizes the balance between detection resolution and phase wrapping frequency.
Solution Approach 2:
The system uses feedback from comparing phase information obtained from patterns with different pitches to determine the actual depth position. By analyzing the phase difference between the first and second patterns, the system can resolve ambiguity from phase wrapping and accurately determine the measurement target point's position in the depth direction.
3Measurement precision
If multiple patterns with different pitches are projected alternately, then the touch recognition accuracy improves, but the measurement time increases
Solution Approach 1:
The patent uses periodic alternation between first and second patterns with different pitches, capturing images at each time point. This periodic measurement approach enables accurate touch recognition through phase comparison while maintaining efficient measurement speed by systematically switching between patterns in a rhythmic manner.
Solution Approach 2:
The system performs preliminary projection of the first pattern followed by the second pattern in sequence, allowing pre-processing of phase information. This preliminary action structure enables efficient computation of depth positions by preparing phase data in a systematic order, reducing overall measurement time while maintaining high touch recognition accuracy.
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 solution achieves high accuracy in touch recognition by reducing phase wrapping and improving depth analysis, allowing for precise detection of finger contact with the screen.
Implementation Method 1
a position detection section adapted to detect a position of the measurement target point with respect to the image display surface based on an image including the measurement target point and the first pattern and an image including the measurement target point and the second pattern obtained by the imaging device
Data Source
AI summary
An image recognition device, an image recognition method and an image recognition unit capable of performing touch recognition high in accuracy. The image recognition device includes a pattern display section adapted to display a first pattern in which bright lines are arranged with a first pitch, and a second pattern in which bright lines are arranged with a second pitch at respective time points different from each other, a measurement point determination section adapted to detect a finger located between a camera and a screen, and determine a fingertip of the finger, and a position detection section adapted to detect the position of the fingertip with respect to the screen based on an image including the first pattern and an image including the second pattern obtained by the camera.