Coordinate Conversion for Accurate OCR on Curved Surfaces
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional optical character recognition (OCR) technologies struggle to accurately recognize characters on curved or inclined surfaces due to distortion, which hinders effective character recognition in images captured by vehicle-mounted cameras.
Innovation Solution
An electronic apparatus equipped with a character recognition section that performs coordinate conversion on image data, utilizing a distance measurement section to measure distances via phase difference detection, and an irradiation section that uses pulsed or diffused light for accurate shape estimation and correction, allowing for precise recognition of characters on non-planar surfaces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional OCR is used to recognize characters on curved or inclined surfaces, then the recognition process can be performed, but the recognition accuracy deteriorates due to character distortion
Solution Approach 1:
The patent introduces a distance measurement dimension by using a distance sensor (Time of Flight system) to measure the distance from the imaging apparatus to various points on the object. This third dimension (depth/distance) information is combined with the two-dimensional image data to reconstruct the three-dimensional shape of the object, enabling accurate coordinate conversion even when the object surface is curved or inclined.
Solution Approach 2:
The patent changes the coordinate system parameters by performing coordinate conversion based on the measured three-dimensional shape information. The conversion uses parameters such as distance, inclination angle, and curvature to transform distorted character coordinates into undistorted reference coordinates, thereby eliminating the distortion problem without changing the physical object.
2Reliability
If a distance measurement section is added to measure distances for coordinate conversion, then character recognition accuracy on non-planar surfaces is improved, but the device complexity increases
Solution Approach 1:
The imaging section is designed to perform multiple functions: it serves as both the imaging device for capturing character images and the distance measurement device for measuring distances to the object. By making the imaging section multi-functional, the patent avoids adding separate dedicated distance measurement hardware, thereby reducing device complexity while still achieving the needed distance information for coordinate conversion.
3Reliability
If coordinate conversion is performed using three-dimensional shape information, then distortion is eliminated and recognition accuracy is improved, but the processing time and computational complexity increase
Solution Approach 1:
The patent performs preliminary action by measuring the three-dimensional shape information and calculating the coordinate conversion parameters before the actual character recognition process. By pre-computing the conversion based on distance and shape data, the system prepares the undistorted coordinate framework in advance, which speeds up the subsequent character recognition step rather than slowing it down.
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 enables accurate recognition of characters on curved or inclined surfaces by converting coordinates to a reference plane, eliminating distortion and enhancing the accuracy of OCR processes.
Implementation Method 1
a distance measurement section that measures distances from a phase difference between reflected light of the irradiation light and the irradiation light
Implementation Method 2
capturing a portion of the electromagnetic radiation reflected by an object
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In an electronic apparatus that captures images, characters in an image are accurately recognized. The electronic apparatus includes an imaging section, a distance measurement section, a shape estimation section, and a coordinate conversion section. The imaging section images an object and captures image data. The distance measurement section measures distances from the imaging section to a plurality of measurement points on a surface of the object. The shape estimation section estimates a shape of the object from the measured distances. The coordinate conversion section performs coordinate conversion on the image data, the coordinate conversion including converting three-dimensional coordinates on the surface of the object into plane coordinates on a predetermined reference plane on the basis of the estimated shape.