基于OpenCV的4K HDMI相机上的自动寻边体系的研究  

Automatic edge detecting in 4K HDMI cameras based on OpenCV

作　　者：刘忠瑞 赵廷玉[1] LIU Zhongrui;ZHAO Tingyu(School of Science,Zhejiang Sci-Tech University,Hangzhou 310018,China)

机构地区：[1]浙江理工大学理学院,杭州310018

出　　处：《智能计算机与应用》2024年第11期156-162,共7页Intelligent Computer and Applications

摘　　要：目前诸如HDMI显微相机等嵌入式设备主要采用肉眼观察+鼠标定位的方式选取物体边缘,存在精度差、耗时久等问题。本文提出一种基于OpenCV视觉库的实时显微图像自动寻边体系。首先,从后台采集实时图像,分离出灰度信息并保存于内存缓冲区;然后,截取鼠标点击处为中心的ROI区域,并基于直方图梯度幅度自动确定Canny边缘检测双阈值获取ROI区域边缘信息;最终,根据不同边缘类型计算鼠标点击处最近的边缘。实验结果表明:相较于肉眼观察+鼠标定位的传统边缘测量方式,本文提出的鼠标定位+自动寻边方式平均耗时缩短约36%,精度从69.86%提升至98.59%,具有精度高、效率高和延时低的特点,在显微测量等领域具有实际应用价值。The method of naked-eye observation-mouse positioning is the main way to detect the edge of objects in the embedded devices represented by HDMI microscope cameras.This conventional edge-detection method inevitably has some problems such as poor precision and long time consuming because it is completely dependent on manual operation.This paper proposes a real-time automatic edge detection system for microscopic images based on the OpenCV vision library.Firstly,a gray-scale image,separated from a real-time mocroscopic image,is stored in the memory buffer.Next,the square area of the gray-scale image centered on the mouse click is truncated as the region of interest(ROI).Then,dual thresholds are determined automatically using Canny edge detection of ROI based on histogram of gradient.Finally,the thresholds are applied to return the edge closest to the mouse-click point according to the different types of edges.The experimental results show that the edge detection method proposed in this paper has the advantages of high precision and short time consuming.Specifically,the average time is shortened by about 36%,while the accuracy is increased from 69.86%to 98.59%compared to the conventional method.It is believed that the automatic edge detection method has a wide application prospect in the field of microscopic detection.

关 键 词：OPENCV 高清多媒体接口 工业相机 自动寻边 边缘检测 

分 类 号：TP311.1[自动化与计算机技术—计算机软件与理论]