Handheld Scanner Image Correction via Optical Speed Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Hand-held scanning devices often produce distorted images due to varying scanning speeds and user operation, leading to low OCR accuracy despite existing solutions that fail to fully correct these distortions.
Innovation Solution
A method involving image binarization, cropping, central line identification, and straightening, combined with variable speed correction, to preprocess images for improved OCR accuracy, utilizing a hand-held scanning device with an optical sensor, processor, and storage for efficient image processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If hand-held scanning device is operated by hand to acquire images, then ease of operation is improved, but image distortion increases due to varying scanning speeds
Solution Approach 1:
The patent replaces mechanical speed measurement systems (rubber rollers, wheels) with an optical-based solution. The method uses the optical sensor already present in the hand-held scanning device to detect text lines and calculate their orientations, thereby determining scanning speed variations without additional mechanical components. This substitution maintains ease of hand-operated scanning while eliminating the complexity and inaccuracy of mechanical speed measurement systems.
Solution Approach 2:
The patent enables the hand-held scanning device to use its own optical sensor for dual purposes: both for acquiring the image data and for measuring the scanning speed. The same sensor that captures the text image is used to detect text line orientations and calculate instantaneous scanning speed. This self-service approach eliminates the need for separate speed measurement mechanisms while maintaining accurate distortion correction.
2Measurement precision
If mechanical structures like rubber rollers are used to measure scanning speed, then speed measurement capability is improved, but device complexity increases
Solution Approach 1:
The patent replaces mechanical speed measurement structures (rubber rollers, wheels, gears) with a computational method using optical sensor data. The system calculates scanning speed by analyzing the orientations of text lines detected by the optical sensor, eliminating the need for additional mechanical components. This reduces device complexity while maintaining speed measurement capability.
Solution Approach 2:
The patent makes the optical sensor multi-functional by using it for both image acquisition and speed measurement. The same sensor that captures the text image is also used to detect text line orientations and calculate scanning speed variations. This universal use of the optical sensor eliminates the need for dedicated speed measurement hardware, thereby reducing device complexity.
3Manufacturing precision
If existing correction methods using character height and font ratio are applied, then some distortion correction is achieved, but OCR accuracy remains too low due to additional user-induced distortions
Solution Approach 1:
The patent performs preliminary detection of text line orientations and calculation of scanning speed variations before applying distortion correction. By first identifying the actual scanning path through orientation analysis and then using this information to guide the correction process, the system addresses both mechanical speed variations and user-induced distortions in a sequential manner, achieving superior OCR accuracy compared to existing methods.
Solution Approach 2:
The patent implements a feedback mechanism where the detected text line orientations are used to calculate correction factors that are then applied to the image. The system continuously adjusts the correction based on the actual orientations detected, creating a closed-loop correction process that adapts to varying scanning conditions and user-induced distortions, thereby improving OCR accuracy.
Data Source
AI summary
A method for correcting an image acquired by a hand-held scanning device. A binarized image of an acquired image is cropped by removing columns on the left end and on the right end of only first components. A work image is created from the cropped image by replacing in each row of components series of first components smaller than a predetermined distance with series of second components. In the work image, a central line is identified. The identified central line in the work image is used to identify the corresponding central line in the cropped image forming a central line image, and in the central line image, the central text line is straightened.