Line-Scan Image Stitching via Lookup Tables for Barcode Scanning
Solution Overview
Problem
Conventional image stitching in barcode scanning and imaging systems is computationally time-consuming and expensive, limiting high-speed object processing, especially in systems using multiple cameras to capture large objects.
Innovation Solution
The use of a lookup table with relative position data between image subsets captured by multiple cameras or image sensors, allowing for image alignment and stitching without processing imaged features, thereby reducing computational requirements and costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional image processing is used to match features and align images from multiple cameras, then image stitching accuracy is improved, but processing time increases and speed decreases
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing transformation parameters (translation, rotation, scaling) in lookup tables during a calibration phase. These parameters are determined beforehand based on known geometric relationships between cameras and the conveyor belt, eliminating the need for real-time feature matching during actual stitching operations. This allows instant retrieval of alignment parameters, achieving both high accuracy and high speed.
2Measurement precision
If high-resolution line scan cameras are used to capture entire items, then image quality and coverage are improved, but system cost increases
Solution Approach 1:
The patent applies segmentation by dividing the field of view into multiple overlapping regions, each captured by a separate camera or camera subset. Instead of using one expensive high-resolution camera to capture the entire conveyor belt width, multiple lower-resolution cameras capture segments that are then stitched together using the lookup table methodology. This reduces individual camera costs while maintaining overall system capability.
Solution Approach 2:
The patent combines multiple lower-resolution images into a single high-resolution composite image through the stitching process. By merging data from multiple cameras with known geometric relationships, the system achieves coverage and quality equivalent to or exceeding a single high-resolution camera, but at lower component costs.
3Area of stationary object
If multiple cameras are used to capture different portions of the scene, then coverage area is improved, but computational complexity increases
Solution Approach 1:
The patent reduces computational complexity by performing all complex geometric calculations in advance during system calibration. The lookup tables store pre-computed transformation parameters that simplify the runtime stitching operation to basic image registration and blending, significantly reducing computational requirements during actual use.
Data Source
AI summary
One embodiment of a system and method for imaging items may include storing at least one lookup table that includes position data of multiple image regions, where the position data provides relative positions of the image regions. An item may be moved to pass through the image regions. Image data of the item may be imaged as the item passes through a plurality of the image regions. The position data may be applied to the captured respective image data from each of the image regions through which the item passed to form a set of stitching image data that is substantially aligned with one another. The set of stitching image data may be stitched to form a composite image of the item.