两种优化小波阈值算法对紫外可见光谱的水质COD数据去噪研究

Study on Denoising of UV-VIS Spectrum Water Quality COD Data by Two Optimized Wavelet Threshold Algorithms

本文主要检测水质COD数据,对数据进行去噪处理,为建立预测水质COD模型提供数据预处理奠定基础。紫外可见光谱检测水质COD数据系统易产生光学仪器本身和外界散射光的噪声影响,得到光谱数据存在不必要的噪声。为开展去噪效果评价,实验稀释了COD浓度为20 mg/L的邻苯二甲酸氢钾溶液来得到了30组水质COD光谱数据,将六种算法去噪效果相互对比,得出灰狼优化小波阈值算法更适合用于去除水质COD数据噪声的结论。为了验证两种优化算法去噪的可行性,采集某内陆河地表水质COD光谱数据进行实例去噪分析。比较两种优化算法的去噪参数评价,灰狼优化小波阈值算法能更好抑制了噪声和提高了系统检测精度,为紫外可见光谱法的水质COD数据的去噪处理提供了一种全新的解决办法,证明了两种优化小波阈值算法适用于地表水质COD数据去噪分析的可行性。

This paper mainly detects the COD data of water quality and denoises the data, which lays a foundation for data preprocessing to establish the COD model for predicting water quality. The UV-VIS spectrum detection water quality COD data system is easy to produce the noise influence of the optical instrument itself and the external scattered light, and there is unnecessary noise in the spectral data. In order to evaluate the denoising effect, 30 groups of water quality COD spectral data were obtained by diluting the potassium hydrogen phthalate solution with COD concentration of 20 mg/L. The denoising effects of the six algorithms were compared with each other, and it was concluded that the gray wolf optimized wavelet threshold algorithm was more suitable for removing the noise of water quality COD data. In order to verify the feasibility of the two optimization algorithms, COD spectral data of surface water quality of an inland river were collected for example denoising analysis. Comparing the denoising parameter evaluation of the two optimization algorithms, the gray wolf optimized wavelet threshold algorithm can better suppress the noise and improve the detection accuracy of the system. It provides a new solution for the denoising of water quality COD data by UV-VIS spectroscopy, and proves that the two optimized wavelet threshold algorithms are suitable for the denoising analysis of surface water quality COD data.