Optical Flow Tracking for Barcode Identification Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Barcode readers often fail to accurately identify items, mistaking similar barcodes as identical or a single barcode as multiple, leading to inaccuracies in item identification.
Innovation Solution
A method using an imaging system with an optical assembly to track barcodes by receiving a series of images, decoding barcodes, identifying key-points, calculating optical flow, and predicting the location of barcodes in subsequent images, allowing for accurate tracking and differentiation between similar codes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional barcode reading methods are used, then the reading process is simple and fast, but the identification accuracy is low and similar barcodes are mistaken as identical
Solution Approach 1:
The barcode is segmented into multiple key-points (corner points, edge points, intersection points) for tracking. Instead of treating the barcode as a single entity, the system divides it into characteristic points that can be individually tracked across frames to distinguish similar barcodes and prevent misidentification.
Solution Approach 2:
The system transitions from 2D barcode pattern recognition to 3D spatiotemporal tracking by adding the time dimension. Optical flow calculations track key-point movements across multiple frames, creating a temporal dimension that distinguishes between similar barcodes based on their motion trajectories rather than just their visual patterns.
2Reliability
If optical flow calculation with multiple key-points is implemented, then tracking accuracy improves, but computational complexity and processing time increase
Solution Approach 1:
The system extracts only the most critical key-points (corner points, edge points, intersection points) from the barcode rather than processing all pixels or features. This selective extraction reduces the number of points requiring optical flow calculation while maintaining sufficient tracking accuracy to distinguish between similar barcodes.
Solution Approach 2:
The system performs preliminary detection and classification of key-point types (corner, edge, intersection) in advance of optical flow calculation. By pre-identifying and categorizing key-points before tracking, the system optimizes the subsequent optical flow computation by applying appropriate algorithms to each key-point type, reducing overall processing time.
3Productivity
If multiple barcodes are tracked simultaneously, then comprehensive monitoring is achieved, but computational load and processing complexity increase
Solution Approach 1:
Each barcode is segmented into its own set of key-points that are independently tracked. The system maintains separate key-point trajectories for multiple barcodes, allowing simultaneous tracking without confusion. This segmentation enables the system to handle multiple barcodes by treating each as an independent collection of trackable features.
Solution Approach 2:
The system creates and maintains separate optical flow models for each tracked barcode. By copying the key-point identification and tracking methodology for each barcode instance, the system can simultaneously monitor multiple barcodes using the same algorithmic framework, reducing the complexity of implementing multi-object tracking from scratch.
Data Source
AI summary
Methods and apparatuses for optical flow estimation for 1D/2D decoding improvements are disclosed herein. An example method includes receiving, from the optical imaging assembly, a series of images including at least a first image and a second image captured over the FOV; decoding a barcode in the first image; identifying a first position of a key-point within the first image; identifying a second position of the key-point within the second image; calculating an optical flow for the barcode based on at least the first position and the second position; and tracking the barcode based on the optical flow.