Prediction of Tidal Flat Sediment Moisture Content Based on Wavelet Transform

The distribution of water in flat tidal sediments will change greatly in space and time, and the changes will lead to the changes of biogenic elements in sediments. Therefore, the tidal flat sediment water content data are monitored in real time, accurately and quickly, which is of great significance to understanding the tidal flat characteristics, grasp the information of tidal flat biogenic elements, and develop tidal flat resources. This paper collected 115 samples of intertidal sediments from Dongdayang village, Qingdao city. The visible near-infrared spectra and moisture content of fresh samples, air-dried for 4 weeks and 8 weeks were measured. The db10 and sym6 wavelet basis were used to transform the original spectrum, and partial least squares regression was used to establish the tidal flat sediment moisture content model. The low-frequency information An and high-frequency information D-n (n=1, 2, ..., 10) of the original spectrum were obtained by 10 order wavelet transform. S- D-n was calculated by the difference between the original spectrum S and D-n. The moisture content models were established using A(n), D-n and S- D-n, respectively, and the results were analyzed. The original spectrum model's R-P(2), RMSEP and RPD were 0. 841, 2. 767 and 2. 481. By analysing low-frequency and high-frequency information, after db10 wavelet basis transforms, the useless information was mainly concentrated in D-3 and D-4, and the accuracy of the moisture content model established by removing D-3 and D-4 was significantly improved, R-P(2) was 0. 878, RMSEP was 2. 501, RPD was 2. 749. Through the analysis of sym6 wavelet basis transform, the useless information was mainly concentrated in D-5 and D-9, the R-P(2), RMSEP and RPD by removing D-5 and D-9 were 0. 87, 2. 475 and 2. 768. Therefore, by analyzing the low-frequency and high-frequency information using wavelet transform, the interference information of sediment moisture content can be effectively found, and the feature information can be extracted. The more accurate the tidal flat sediment moisture content model is established, it provides a theoretical basis for real-time and dynamic monitoring of tidal flat sediment moisture content.