摘要
针对“机器视觉”课程实验案例的需要,设计了八音盒视频运动音符识别实验方案。八音盒依靠发条储存的动力匀速转动滚轴,通过滚轴上不同位置的凸起对簧片拨动发出旋律。提出一种基于机器视觉方式提取并恢复八音盒乐谱特征的方案。该方案采用图像形态学、机器视觉、音频信号的时频分析等研究方法,成功提取转动滚轴的运动特征并与音符特征像匹配,结合C++编程和乐器数字接口编程,实现运行的八音盒视频实时音符特征提取,将提取的特征输出并实现实时演奏。该课程实验培养学生独立自主的思考和设计能力、综合各类知识的实践能力。
Machine vision is a practical course.In view of the course experiment case,this paper designs an experimental scheme of eight-tone box video motion notes recognition.The eight-tone box relies on the power stored in the hair bar to rotate the roller at a uniform speed,and the spring is rotated by the bulges at different positions on the roller.In this paper,a method based on machine vision is proposed to extract and restore eight-tone box music features.In this scheme,image morphology,visual measurement,time-frequency analysis of audio signal and other research methods are used to successfully extract the motion characteristics of the rotating roller and match the feature image of the notes.Combined with C++ programming and musical instrument digital interface(MIDI) programming,the real-time feature extraction of the running eight-tone box video notes is realized,and the extracted features are output and real-time performance is realized.This course has cultivated the students' independent thinking and design ability as well as the practical ability of synthesizing all kinds of knowledge.
作者
于蕾
张宇
柳守一
汲清波
YU Lei;ZHANG Yu;LIU Shouyi;JI Qingbo(College of Information and Communication Engineering,Harbin Engineering University,Harbin 150001,China)
出处
《实验室研究与探索》
CAS
北大核心
2023年第5期149-152,196,共5页
Research and Exploration In Laboratory
基金
2021年黑龙江省教改项目(SJGY20210167,SJGY20210215)
2022年哈尔滨工程大学教育教学改革项目(JY2022Y034)。
关键词
图像处理
图像形态学
时频分析
MIDI编程
image processing
image morphology
time-frequency analysis
MIDI programming
