南音工乂谱图像预处理的流程和方法
|
摘要
分解南音工乂谱的谱式为谱字、骨指符号和撩拍符号等3个部分。建立南音工乂谱图像的预处理流程,包括灰度化、清晰度计算、去噪、增强、扩大、二值化和细化等。其中,必要流程包括灰度化、增强、二值化和细化等。根据工乂谱图像的质量、大小等的属性不同,以处理规范和需要为目标,可以选择清晰度计算、去噪、扩大等流程。以中等质量的白底黑字的乐谱图像为样谱,利用MATLAB对预处理各个过程的方法进行了最优化实验。根据实验记录和结果,比较可供选择的多种算法,列出了乐谱图像进行预处理实验效果最好的算法、主要参数和建议,为乐谱图像的进一步分割和识别等准备条件。
The decomposition of the Nanyin Gongyi spectrum is made up of three parts: spectrum characters, bone finger symbols and beat symbols. The preprocessing flow of Nanyin Gongyi spectrum image is established, including gray level, definition calculation, denoising, enhancement, enlargement, binarization and thinning, etc. Among them, the necessary processes include grayscale, enhancement, binarization and refinement, etc. According to the different attributes of the image quality, size and so on, aiming at processing specification and needs, clearness calculation, denoising and enlargement can be selected. Using medium-quality white background and black characters as samples, based on MATLAB to optimize the pretreatment methods of each process. According to the experimental records and comparing with alternative algorithms, the best algorithm, main parameters and conclusions of music score image preprocessing experiment are listed, provide basic conditions for subsequent segmentation and recognition of spectrum.
The decomposition of the Nanyin Gongyi spectrum is made up of three parts: spectrum characters, bone finger symbols and beat symbols. The preprocessing flow of Nanyin Gongyi spectrum image is established, including gray level, definition calculation, denoising, enhancement, enlargement, binarization and thinning, etc. Among them, the necessary processes include grayscale, enhancement, binarization and refinement, etc. According to the different attributes of the image quality, size and so on, aiming at processing specification and needs, clearness calculation, denoising and enlargement can be selected. Using medium-quality white background and black characters as samples, based on MATLAB to optimize the pretreatment methods of each process. According to the experimental records and comparing with alternative algorithms, the best algorithm, main parameters and conclusions of music score image preprocessing experiment are listed, provide basic conditions for subsequent segmentation and recognition of spectrum.
引文
[ 1 ]刘佳.学科系统化与研究精细化的产物:论南音学形成的学术轨迹[J].音乐艺术,2016(2):39-43.
[ 2 ]吴雅蓉.试论南音工乂谱解读与弹奏的若干“盲区”[J].泉州师范学院学报,2016,34(5):98-101.
[ 3 ]曾台盛,朱达欣.南音工乂谱软件设计与实现[J].福建电脑,2018(11):50-52.
[ 4 ]曾宪林.南音本体音乐分析理论与方法构想[J].中国音乐学,2015(2):91-99.
[ 5 ]陈亮,李卫军,谌琛.数字图像清晰度评价函数的通用评价能力研究[J].计算机工程与应用,2013,49(14):152-155.
[ 6 ]张岩,吕梦儒.基于MATLAB的音乐旋律二维可视化方法[J].沈阳师范大学学报(自然科学版),2018,36(4):342-346.
[ 7 ]李宸鑫.基于MATLAB三种滤波算法的图像去噪技术研究[J].通讯世界,2018(6):283-284.
[ 8 ]陈洁萍.手写签名脱机识别图像的预处理方法[J].轻工科技,2013(7):92-163.
[ 9 ]张玉薇,杨建斌.一种便捷数字图像处理系统的MATLAB实现[J].信息记录材料,2019,20(2):40-42.
[10]于岗,熊建设.基于扩大采样范围的图像放大算法[J].电子制作,2015(5):55-56.
[11]王研,张岩.基于Hadoop的云平台的实现与基准测试[J].沈阳师范大学学报(自然科学版),2016,34(2):240-245.
[12]许秦蓉.EMS表单中手写中文识别图像预处理方法研究[J].包装工程,2014,35(21):80-84.
[13]张光建.手写数字识别的图像预处理GUI设计[J].智能计算机与应用,2015,5(1):98-100.
[14]叶福玲.一种改进的图像骨架提取算法[J].西昌学院学报(自然科学版),2018,32(3):91-94.
[15]白天毅.基于神经网络的手写体数字识别关键技术研究[D].西安:西北工业大学,2014.
[ 2 ]吴雅蓉.试论南音工乂谱解读与弹奏的若干“盲区”[J].泉州师范学院学报,2016,34(5):98-101.
[ 3 ]曾台盛,朱达欣.南音工乂谱软件设计与实现[J].福建电脑,2018(11):50-52.
[ 4 ]曾宪林.南音本体音乐分析理论与方法构想[J].中国音乐学,2015(2):91-99.
[ 5 ]陈亮,李卫军,谌琛.数字图像清晰度评价函数的通用评价能力研究[J].计算机工程与应用,2013,49(14):152-155.
[ 6 ]张岩,吕梦儒.基于MATLAB的音乐旋律二维可视化方法[J].沈阳师范大学学报(自然科学版),2018,36(4):342-346.
[ 7 ]李宸鑫.基于MATLAB三种滤波算法的图像去噪技术研究[J].通讯世界,2018(6):283-284.
[ 8 ]陈洁萍.手写签名脱机识别图像的预处理方法[J].轻工科技,2013(7):92-163.
[ 9 ]张玉薇,杨建斌.一种便捷数字图像处理系统的MATLAB实现[J].信息记录材料,2019,20(2):40-42.
[10]于岗,熊建设.基于扩大采样范围的图像放大算法[J].电子制作,2015(5):55-56.
[11]王研,张岩.基于Hadoop的云平台的实现与基准测试[J].沈阳师范大学学报(自然科学版),2016,34(2):240-245.
[12]许秦蓉.EMS表单中手写中文识别图像预处理方法研究[J].包装工程,2014,35(21):80-84.
[13]张光建.手写数字识别的图像预处理GUI设计[J].智能计算机与应用,2015,5(1):98-100.
[14]叶福玲.一种改进的图像骨架提取算法[J].西昌学院学报(自然科学版),2018,32(3):91-94.
[15]白天毅.基于神经网络的手写体数字识别关键技术研究[D].西安:西北工业大学,2014.