Bar code recognition method and bar code recognition device for a code scanning terminal, and code scanning terminal
By integrating an NPU unit and a probabilistic inference model into the barcode scanning terminal, multi-frame barcode images are fused and their solvability is evaluated. Appropriate decoding paths are allocated, which solves the CPU usage problem caused by invalid computation and improves the response speed and efficiency of the barcode scanning terminal.
CN122414218APending Publication Date: 2026-07-17BEIJING TSINGTENG MICROSYSTEM CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIJING TSINGTENG MICROSYSTEM CO LTD
- Filing Date
- 2026-04-02
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing barcode scanning terminals suffer from a large amount of invalid computation when processing low-quality barcode images, resulting in high CPU usage and slow response speed.
Method used
An NPU unit is used to integrate a probabilistic inference model. Multi-channel input tensors are generated by fusing multiple frames of grayscale barcode images. The probabilistic inference model is used to determine the solvability probability of each frame image, and fast or slow decoding paths are allocated for decoding based on the solvability probability.
Benefits of technology
This reduces the CPU usage of invalid calculations and improves the response speed and efficiency of barcode recognition.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN122414218A_ABST
Abstract
本申请涉及图像识别技术领域，公开了一种用于扫码终端的条码识别方法及条码识别装置、扫码终端。扫码终端集成有NPU单元；用于扫码终端的条码识别方法，包括：采集并缓存连续的多帧灰度条码图像；将相邻多帧灰度条码图像进行融合，生成多通道输入张量；利用NPU单元部署的概率推理模型对多通道输入张量进行推理，确定每帧灰度条码图像的可解性概率；根据可解性概率确定多帧灰度条码图像中能够解码的目标帧条码图像，并按照可解性概率所处的区间范围，为目标帧条码图像分配快速解码路径或者慢速解码路径执行条码解码。本申请可以减少解码时无效计算对扫码终端CPU单元的占用，提高条码识别的响应速度。
Need to check novelty before this filing date? Find Prior Art