Marker Detection Using Multiple Reference Positions for AR Display Stability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The visibility of display information in augmented reality image processing is compromised due to quantization errors and imaging blur, leading to unstable display positions and reduced accuracy in marker detection, which affects the precision of translation and rotation vectors used for superimposing information on images.
Innovation Solution
The image processing apparatus improves visibility by using a method that includes recognizing marker IDs with sub-pixel accuracy, employing interpolation and priority ordering of shape features, and calculating translation and rotation vectors based on multiple reference positions to reduce errors caused by quantization and imaging blur, thereby stabilizing the display of information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If marker detection is performed using four corners of the shot marker, then the display position can be determined, but quantization errors and imaging blur cause unstable display positions and reduced accuracy
Solution Approach 1:
The patent segments the marker detection process into multiple independent reference position detections rather than relying on a single four-corner detection. By dividing the marker into multiple reference positions and detecting each independently, the system reduces the impact of quantization errors and imaging blur on overall detection accuracy.
Solution Approach 2:
The patent uses more reference positions than the minimum required (excessive action) to compensate for errors. By detecting multiple reference positions and using their average or weighted combination, the system achieves higher reliability than would be possible with the standard four-corner method alone.
2Productivity
If translation and rotation vectors are calculated from four corners of the marker, then the superposition can be performed, but errors in vector calculation reduce the precision of display information
Solution Approach 1:
The patent segments the vector calculation process by computing translation and rotation vectors from multiple reference position pairs rather than from a single four-corner configuration. This segmentation allows error distribution and reduces the impact of individual measurement errors on the final vector accuracy.
Solution Approach 2:
The patent performs preliminary detection of multiple reference positions and pre-calculates their coordinates before computing the final translation and rotation vectors. This preliminary action allows for error analysis and correction before the actual vector calculation, improving overall precision.
3Device complexity
If display information is superposed based on reference position from four corners, then the process is simple, but the visibility and stability of display information are compromised
Solution Approach 1:
The patent segments the display information superposition process into multiple stages: detecting multiple reference positions, calculating their average position, and then superposing based on this stabilized reference. This segmentation maintains reasonable complexity while significantly improving visibility and stability.
Solution Approach 2:
The patent introduces an intermediary calculation step that computes the average position of multiple reference points before using it for superposition. This intermediary acts as a mediator that filters out quantization errors and imaging blur effects, providing a more stable basis for display information placement.
Data Source
Figure 1
Figure 2A~2C
Figure 3
AI summary
An object of the disclosed technique is to improve the visibility to the user of the display information. An image processing apparatus includes an obtaining unit that obtains an image including a marker that specifies display information and a reference position for displaying the display information, from an image pattern including a first pixel area and a second pixel area; an extraction unit that extracts an intersecting point of a first line segment and a second line segment, the first line segment being specified based on a first boundary and a second boundary, the first boundary being a boundary in a first connected area where the first pixel area and the second pixel area are connected in the vertical direction to each other, the second boundary being a boundary between the first pixel area and the second pixel area respectively point-symmetric with respect to the first connected area, the second line segment being specified based on a third boundary and a fourth boundary, the third boundary being a boundary in a second connected area where the first pixel area and the second pixel area are connected in the horizontal direction to each other, the fourth boundary being a boundary between the first pixel area and the second pixel area respectively point-symmetric with respect to the second connected area, a first apex of a four-cornered shape of an outer edge of the marker including the first pixel area, or a plurality of apexes of the first pixel area, or a second apex of a plurality of second pixel areas; and a calculation unit that calculate a display position of the display information using the first apex, the second apex or the intersecting point as the reference position.