基于中值滤波去噪算法的激光光谱图像数据库构建  

Construction of laser spectral image database based on median filter denoising algorithm

作　　者：王研 梁洪涛 杨阿荣 WANG Yan;LIANG Hongtao;YANG Arong(Department of Accounting Information,Miyun Campus of Capital University of Economics and Business,Beijing 101500,China)

机构地区：[1]首都经济贸易大学密云分校,北京101500

出　　处：《激光杂志》2022年第2期110-113,共4页Laser Journal

基　　金：教育部产学合作协同育人项目(No.783737c3-3976-41b9-9cc1-ae334fce2435)。

摘　　要：在目前建立的激光光谱图像数据库中,数据库的滤波性能较差,导致存储的光谱图像数据受噪声影响严重。因此,提出基于中值滤波去噪算法的激光光谱图像数据库构建。首先利用多数据源数据库模式,建立数据库的存储引擎。其次利用中值滤波去噪算法,在数据库中添加去噪模块。然后将滤波后的图像进行光学参数转换,最后对不同光谱数据分别进行格式转换,并对数据进行归一化存储至数据库中,完成激光光谱图像数据库设计。利用不同方法的图像数据库与设计数据库对生成的激光光谱数据进行存储,并进行对比实验,分别对数据库的构建以及滤波能力进行分析。实验结果证明设计的激光光谱图像数据库可以实现正常的数据库构建,具有较好的滤波性能,满足设计初衷。In current laser spectral image database, the filtering performance of database is poor, resulting in spectral image data stored seriously affected by noise. Therefore, construction of laser spectral image database based on median filter denoising algorithm is proposed. Firstly, the storage engine of database is established by using multi-data source database mode. Secondly, median filter denoising algorithm is used to add denoising module in database. Then, the filtered image is converted into optical parameters. Finally, different spectral data are converted into different formats, and the data are normalized and stored in database to complete the design of laser spectral image database. Different methods of image database and design database are used to store the generated laser spectrum data, and comparative experiments are carried out to analyze the construction of database and filtering ability. The experimental results show that the designed laser spectral image database can achieve normal database construction, has good filtering performance, and meets original design intention.

关 键 词：中值滤波去噪算法 图像数据库 滤波性能 去噪模块 归一化存储 

分 类 号：TN297[电子电信—物理电子学]