Research on Kalman Spectral Filtering Method Using Improved PSO Adaptive Optimization

HOU Maosheng; SHAO Yuheng; LIU Tao; LIU Chuang; WENG Mingwang; WANG Tongyu

doi:10.3788/gzxb20265508.0830002

Chinese

您当前的位置：

首页 >

文章列表页 >

Research on Kalman Spectral Filtering Method Using Improved PSO Adaptive Optimization

更新时间：2026-07-15

- Research on Kalman Spectral Filtering Method Using Improved PSO Adaptive Optimization
- Acta Photonica Sinica Pages: 1-13(2026)
- 作者机构：
  
  长春理工大学光电工程学院，光电测控与光信息传输技术教育部重点实验室，长春 130022
- 作者简介：
- 基金信息：
  
  Jilin Province Technology Innovation Guidance Project(20230401101YY);Key Projects of Jilin Provincial Department of Education(JJKH20240925KJ)
- DOI：10.3788/gzxb20265508.0830002
  CLC： O433.4
- CSTR：32255.14.gzxb20265508.0830002
- Received：07 April 2026，
  
  Revised：2026-05-13，
  
  Accepted：11 June 2026，
- 稿件说明：
移动端阅览
侯茂盛，邵宇衡，刘涛，等. 结合改进PSO自适应寻优的卡尔曼光谱滤波方法研究［J］.光子学报，2026，55（8）：0830002

HOU Maosheng， SHAO Yuheng， LIU Tao， et al. Research on Kalman Spectral Filtering Method Using Improved PSO Adaptive Optimization［J］. Acta Photonica Sinica， 2026， 55（8）：0830002
侯茂盛，邵宇衡，刘涛，等. 结合改进PSO自适应寻优的卡尔曼光谱滤波方法研究［J］.光子学报，2026，55（8）：0830002 DOI： 10.3788/gzxb20265508.0830002. CSTR： 32255.14.gzxb20265508.0830002.

HOU Maosheng， SHAO Yuheng， LIU Tao， et al. Research on Kalman Spectral Filtering Method Using Improved PSO Adaptive Optimization［J］. Acta Photonica Sinica， 2026， 55（8）：0830002 DOI： 10.3788/gzxb20265508.0830002. CSTR： 32255.14.gzxb20265508.0830002.

摘要

针对传统卡尔曼滤波在处理光谱滤波信号时，需要手动调参，且常出现稳定性差、适应性弱及滤波失真等问题，提出一种结合改进粒子群自适应寻优的卡尔曼滤波方法，该方法构建了均方误差和反平滑度相结合的适应度函数以适用于光谱滤波要求，可实现过程激励噪声与测量噪声协方差矩阵的自适应寻优，解决了传统卡尔曼滤波方法需要人工调参的低效率问题，并保证滤波有效性且留存更多波形细节特征。建立了均方误差、信噪比和反平滑度作为滤波方法量化评价指标，分别选取两组模拟光谱数据、无水乙醇和聚丙烯的实测拉曼数据，对所研究方法滤波效果的有效性和适用性进行验证。最后，引入中值滤波、滑动平均滤波和传统卡尔曼滤波与所研究方法的滤波效果进行比较。结果表明，所提出方法的综合性能更优：模拟光谱数据中，研究方法得出的均方误差分别为87.424、3.814，高低噪声场景的信噪比分别达39.03dB、23.2dB，反平滑度贴近原始信号；乙醇与聚丙烯实测数据的均方误差分别为7.789与0.027、信噪比达45.33dB与72.35dB，能有效抑噪且保留光谱细节特征。该方法摆脱了人工调节参数的依赖，提升了光谱滤波的鲁棒性与自动化水平，为光谱数据处理提供了新方案。

Abstract

Spectral filtering algorithms constitute a core component of spectral detection technology. Traditional spectral filtering methods， however， suffer from various shortcomings： median filtering lacks accuracy for complex dynamic signals， moving average filtering has limited noise adaptability， and Kalman filtering is difficult to tune manually. Few adaptive Kalman filtering methods can be directly applied to spectral filtering. To overcome these issues， this study proposes a Kalman spectral filtering method that employs improved Particle Swarm Optimization （PSO） for adaptive parameter optimization. This method aims to overcome the parameter tuning difficulty， insufficient accuracy， and poor stability of traditional Kalman filtering in spectral filtering. It also targets higher accuracy than other traditional filters.In this method， a Kalman filter is integrated with PSO-based adaptive optimization. The particle swarm algorithm innovatively adopts a weighted fitness function that combines Mean Square Error （MSE） and anti-smoothness. This function is specifically designed for spectral filtering. This allows adaptive optimization of the process noise and measurement noise covariance matrices. Such optimization overcomes the inefficiency of manual parameter tuning in traditional Kalman filtering. As a result， effective filtering is achieved while more detailed waveform features are preserved. The effectiveness and applicability of the proposed algorithm are validated using four datasets： two sets of simulated spectral data， a measured Raman spectrum of anhydrous ethanol， and a measured Raman spectrum of polypropylene. Meanwhile， to clearly demonstrate its advantages， the proposed method is compared with three traditional methods. Mean square error， signal-to-noise ratio， and anti-smoothness are used as evaluation indicators.Simulation and experimental results demonstrate that the proposed method achieves superior overall performance. For the two simulated datasets under low-noise and high-noise conditions， the proposed algorithm achieves MSE values of 3.814 and 87.424， respectively. This indicates that the filtering error is greatly reduced， and the algorithm can effectively eliminate noise interference while significantly improving signal quality. In terms of SNR， the proposed method reaches 39.03 dB in the low-noise scenario and 23.2 dB in the high-noise scenario， exhibiting excellent noise suppression capability， far exceeding those of the three traditional methods. In terms of anti-smoothness， the proposed method is the closest to the original spectral signal among all compared methods. This indicates that the algorithm can effectively retain the detailed characteristics of the original spectral signal while suppressing noise. The experimental results of the measured ethanol and polypropylene Raman spectra further demonstrate that the proposed algorithm is practical and reliable. For the filtered data， the MSE values are 7.789 and 0.027， respectively. This demonstrates the high accuracy of the proposed algorithm in practical spectral filtering. The SNR values of the filtered data reach 45.33 dB and 72.35 dB， respectively， both being the highest among all compared methods. This shows that the algorithm has strong adaptability and noise suppression ability in actual measurement scenarios. In addition， the anti-smoothness of the proposed algorithm is also the best among the four filtering methods. This further confirms that the algorithm can well balance two core requirements： strong noise suppression and detailed feature preservation.Therefore， the proposed method has three main advantages. First， it significantly reduces the reliance on manual parameter tuning， solving the associated problems of high tuning difficulty and poor consistency in traditional Kalman filtering. Second， it greatly improves the accuracy and robustness of spectral filtering. Meanwhile， its good performance in anti-smoothness shows that this method balances strong noise suppression and detailed feature preservation well. This addresses the key challenge of traditional filtering methods： balancing noise suppression and detail preservation. In summary， the proposed Kalman spectral filtering algorithm using improved PSO adaptive optimization provides a new and effective method for spectral filtering in various fields. It also offers significant theoretical reference value and promising practical application prospects for the advancement of spectral detection technology.

关键词

Keywords

references

XIAO Qin ， LI Zhengzhou ， LIU Haiyi . Infrared non-uniformity correction method based on scene-adaptive directional guided filtering ［J］. Acta Photonica Sinica ， 2024 ， 53 （ 11 ）： 1110001 .

肖沁，李正周，刘海毅 . 基于场景自适应方向引导滤波的红外成像非均匀性校正方法［J］. 光子学报， 2024 ， 53 （ 11 ）： 1110001 .

GAO Xiangyu ， LI juan ， YU Can ， et al ， Interferogram filtering method for interferometric spectral imager based on VMD-DWT ［J］. Acta Photonica Sinica ， 2025 ， 54 （ 4 ）： 0410001 .

高翔宇，李娟，于粲，等 . 基于VMD-DWT的干涉型光谱成像仪干涉图滤波方法［J］. 光子学报， 2025 ， 54 （ 4 ）： 0410001 .

YANG Yanwei ， HE Xiaojian ， PAN Baowu ， et al . Parameter optimization of laser-induced breakdown bauxite spectra based on cavity confinement ［J］. Infrared and Laser Engineering ， 2022 ， 51 （ 3 ）： 251 - 257 .

杨彦伟，郝晓剑，潘保武，等 . 基于腔体约束激光诱导击穿铝土矿光谱的参数优化［J］. 红外与激光工程， 2022 ， 51 （ 3 ）： 251 - 257 . DOI: 10.3788/IRLA20210661 http://dx.doi.org/10.3788/IRLA20210661

YAN Chunsheng . A review on spectral data preprocessing techniques for machine learning and quantitative analysis ［J］. iScience ， 2025 ， 28 （ 7 ）： 112759 - 112759 .

LU Rongsheng ， ZHANG Zilong ， ZHANG Ailin ， et al . Research progress on key technologies of chromatic confocal sensors （invited）［J］. Laser Optoelectronics Progress ， 2024 ， 61 （ 2 ）： 77 - 104 .

卢荣胜，张紫龙，张艾琳，等 . 光谱共焦传感器关键技术研究进展（特邀）［J］. 激光与光电子学进展， 2024 ， 61 （ 2 ）： 77 - 104 .

ZHONG Minghua ， CHANG Huanqiu . Correction of flame atomic absorption spectroscopy interference using the Kalman filtering method ［J］. Chinese Journal of Analysis Laboratory ， 1996 ，（ 3 ）： 72 - 75 .

衷明华，常焕球 . Kalman滤波法校正火焰原子吸收光谱干扰［J］. 分析试验室， 1996 ，（ 3 ）： 72 - 75 .

LI Suwen ， LIU Wenqing ， XIE Pinhua . Differential optical absorption spectroscopy based on Kalman filter ［J］. Journal of South China University of Technology（Natural Science Edition）， 2006 （ 10 ）： 113 - 116，122 .

李素文，刘文清，谢品华 . 基于卡尔曼滤波的差分光学吸收光谱法［J］. 华南理工大学学报（自然科学版）， 2006 （ 10 ）： 113 - 116，122 .

LI Chunyan ， WANG Ninglin ， LIU Jihong ， et al . Analysis and research of spectral confocal displacement measurement method based on multiple scattering correction ［J］. Infrared and Laser Engineering ， 2025 ， 54 （ 1 ）： 231 - 240 .

李春艳，王泞淋，刘继红，等 . 基于多元散射校正的光谱共焦位移测量方法分析与研究［J］. 红外与激光工程， 2025 ， 54 （ 1 ）： 231 - 240 . DOI: 10.3788/IRLA20240407 http://dx.doi.org/10.3788/IRLA20240407

ZHANG Kai ， ZUO Wangmeng ， CHEN Yunjin ， et al . Beyond a Gaussian denoiser： residual learning of deep CNN for image denoising ［J］. IEEE Transactions on Image Processing ， 2017 ， 26 （ 7 ）： 3142 - 3155 .

YUAN Qiangqiang ， ZHANG Qiang ， LI Jie ， et al . Hyperspectral image denoising employing a spatial–spectral deep residual convolutional neural network ［J］. IEEE Transactions on Geoscience and Remote Sensing ， 2019 ， 57 （ 2 ）： 1205 - 1218 .

JIAO Ting ， DENG Hao ， XU Zhenyu ， et al . Research on dual comb spectroscopy correction algorithm for innovation-based adaptive extended kalman filtering ［J］. Chinese Journal of Lasers ， 2025 ， 52 （ 10 ）： 252 - 263 .

焦婷，邓昊，许振宇，等 . 基于新息自适应扩展卡尔曼滤波的双光梳光谱校正算法研究［J］. 中国激光， 2025 ， 52 （ 10 ）： 252 - 263 .

LU Zhiyi ， LI Xiangping ， CHEN Qi ， et al . Kalman filtering decoupling algorithm based on particle swarm optimization ［J］. Systems Engineering and Electronics ， 2018 ， 40 （ 4 ）： 751 - 755 .

陆志毅，李相平，陈麒，等 . 基于粒子群优化的卡尔曼滤波去耦算法［J］. 系统工程与电子技术， 2018 ， 40 （ 4 ）： 751 - 755 .

ANWAR S ， JAVED I B ， ABDUL W K ， et al . Comparative analysis of median filter and its variants for removal of impulse noise from gray scale images ［J］. Journal of King Saud University - Computer and Information Sciences ， 2022 ， 34 （ 3 ）： 505 - 519 .

SHAN Z ， YANG J ， SANJUAN M A F ， et al . A novel adaptive moving average method for signal denoising in strong noise background ［J］. The European Physical Journal Plus ， 2022 ， 137 ： 50 .

MOHAMED A ， SCHWARZ K ， Adaptive Kalman Filtering for INS/GPS ［J］. Journal of Geodesy ， 1999 ， 73 ： 193 - 203 .

GE Baoshuang ， ZHANG Hai ， TANG Zhikun . Improved robust adaptive Kalman filter with innovation- based outliers diagnosis ［J］. Navigation Positioning and Timing ， 2020 ， 7 （ 1 ）： 48 - 54 .

葛宝爽，张海，唐志坤 . 基于新息异常检测的改进抗差自适应卡尔曼滤波算法［J］. 导航定位与授时， 2020 ， 7 （ 1 ）： 48 - 54 .

LI Dehao ， WANG Dan ， LI Zhiyan ， et al . Application of Kalman filter in gas detection by cavity ring-down spectroscopy ［J］. Spectroscopy and Spectral Analysis ， 2024 ， 44 （ 10 ）： 2727 - 2732 .

李德浩，王丹，李治艳，等 . 卡尔曼滤波方法在腔衰荡光谱技术探测气体中的应用［J］. 光谱学与光谱分析， 2024 ， 44 （ 10 ）： 2727 - 2732 .

CLERC M ， KENNEDY J . The particle swarm - explosion， stability， and convergence in a multidimensional complex space ［J］. IEEE Transactions on Evolutionary Computation ， 2002 ， 6 （ 1 ）： 58 - 73 .

WANG Dongfeng ， MENG Li . Performance analysis and parameter selection of PSO algorithms ［J］. Acta Automatica Sinica ， 2016 ， 42 （ 10 ）： 1552 - 1561 .

王东风，孟丽 . 粒子群优化算法的性能分析和参数选择［J］. 自动化学报， 2016 ， 42 （ 10 ）： 1552 - 1561 .

EILERS P H . A perfect smoother ［J］. Analytical Chemistry ， 2003 ， 75 （ 14 ）： 3631 .

JANACEK G . Time series analysis forecasting and control ［J］. Journal of Time Series Analysis ， 2010 ， 31 （ 4 ）： 303 - 303 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Approach for Measuring the Optimal Local Optical Power of Coherent Detection System （Invited）

Application of Wavelet Threshold Denoising Algorithm in Photodetector Signal Processing

The Design of a Photographic Objective Using Cubic Phase Plate

Research on Feasibility of GEO Target Visual Detection

光学相干层析系统噪音分析(Ⅰ)——理论与计算

Related Author

Yazhou YUE

Wenhua JIN

Hongjie LEI

Bin LI

WU Yan

ZHANG Rong-zhu

杨皓明张新方志良雷广智许英朝张欣张建萍何锋贇

车驰骋李英才陈荣利樊学武马臻

Related Institution

Xi'an Flight Automatic Control Research Institute， AVIC

College of Electronics Information Engineering, Sichuan University

南开大学现代光学研究所光电信息技术教育部重点实验室

中国科学院西安光学精密机械研究所

(1 南开大学现代光学研究所,光电信息技术科学教育部重点实验室, )

AI问答

Postal code：100079
Tel：（010）53879206 Email：tmw@bjxintong.com.cn
Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09082226号-64 京公网安备11010602201714号
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰