CN-121598339-B - Sea plankton density prediction method based on environmental factors and acoustics

CN121598339BCN 121598339 BCN121598339 BCN 121598339BCN-121598339-B

Abstract

The invention discloses a sea area plankton density prediction method based on environmental factors and acoustics, which belongs to the technical field of marine organism prediction and comprises the steps of obtaining multi-source data from a target sea area, wherein the multi-source data comprise acoustic echo data and environmental factor data, extracting noise level from the acoustic echo data, judging whether the noise level exceeds a preset threshold value, denoising the acoustic echo data to obtain a pure acoustic wave signal characteristic set, normalizing characteristics and environmental factor data in the pure acoustic wave signal characteristic set, and obtaining an estimation result of plankton distribution density through a multiple regression equation according to the normalized pure acoustic wave signal characteristic and environmental factor data. The sea area plankton density prediction method based on the environmental factors and the acoustics solves the problem that in the aspect of the current plankton density prediction, the prediction is inaccurate due to the fact that the sea environment is complex.

Inventors

PENG YALAN
LI YAQUAN
ZHANG SHUYONG
WU DI
ZHANG YAN
ZHOU HAODA
SUN MINGSHUAI

Assignees

自然资源部珠海海洋中心(自然资源部珠海海洋预报台)
台山核电合营有限公司
中国水产科学研究院南海水产研究所

Dates

Publication Date: 20260512
Application Date: 20260127

Claims (7)

1. The sea plankton density prediction method based on the environmental factors and the acoustics is characterized by comprising the following steps of: s1, acquiring multi-source data from a target sea area, wherein the multi-source data comprises acoustic echo data and environmental factor data; s2, extracting noise level from the acoustic echo data; in the step S2, the environmental factor data includes chlorophyll a concentration, turbidity and dissolved oxygen, and in the step S2, the chlorophyll a concentration, turbidity and dissolved oxygen with the same time stamp are spliced to obtain spliced data, and the spliced data are combined to obtain an environmental factor set; Performing preliminary classification treatment on the environmental factor set through KMeans clustering of the Scikit-learn library to obtain a preliminary clustering result, generating corresponding pseudo tags according to the preliminary clustering result, and endowing each environmental factor data with corresponding pseudo tags to obtain environmental factor data with pseudo tags; The step of supervised classification by adopting the trained XGBoost comprises the steps of taking the environmental factor data as the input of the trained XGBoost and outputting the environmental factor data and the corresponding labels after classification; acquiring acoustic echo data of the same time stamp based on the classified environmental factor data, and giving labels of the same category as the classified environmental factor data; then extracting the corresponding noise level according to the acoustic echo data of the same tag; s3, judging whether the noise level exceeds a preset threshold value, and if so, denoising the acoustic echo data to obtain a pure acoustic signal feature set; s4, normalizing the characteristics and the environmental factor data in the pure acoustic wave signal characteristic set; s5, obtaining an estimation result of plankton distribution density through a multiple regression equation according to the normalized pure acoustic wave signal characteristics and the environment factor data.
2. A method for predicting sea plankton density based on environmental factors and acoustics as in claim 1, wherein: In the step S3, it is determined whether the noise levels of the acoustic echo data of the different tags exceed corresponding preset thresholds, respectively, where the preset thresholds are determined according to the categories of the acoustic echo data.
3. A method for predicting sea plankton density in a sea area based on environmental factors and acoustics as set forth in claim 2, wherein in said step S2, acoustic echo data of different tags are combined into different acoustic wave feature sets, and the noise level is determined by calculating standard deviation of the acoustic echo data in the different acoustic wave feature sets.
4. The method for predicting sea plankton density in sea area based on environmental factors and acoustics according to claim 3, wherein in step S3, a preset threshold corresponding to the label of the acoustic wave feature set is obtained from a preset threshold library; Judging whether the noise level exceeds a preset threshold value or not to obtain a judging result; and if the judging result is displayed to be exceeded, processing the sound wave characteristic set by adopting a preset denoising method to obtain a pure sound wave signal characteristic set.
5. The method for predicting the sea plankton density in the sea area based on the environmental factors and the acoustics of claim 3, wherein in the step S4, the characteristics, chlorophyll a concentration, turbidity and dissolved oxygen amplitudes of the pure acoustic wave signal characteristic set are respectively adjusted by adopting a normalization method to respectively obtain a normalized pure acoustic wave signal characteristic set, a chlorophyll a concentration set, a turbidity set and a dissolved oxygen set, and the data amplitudes of the normalized pure acoustic wave signal characteristic set, the normalized chlorophyll a concentration set, the normalized turbidity set and the normalized dissolved oxygen set are all within a preset interval [0,1 ].
6. The method for predicting sea plankton density based on environmental factors and acoustics according to claim 5, wherein in the step S5, a linear regression algorithm of sklearn library is adopted to build a prediction model, the prediction model takes a normalized pure sound wave signal characteristic set, chlorophyll a concentration set, turbidity set and dissolved oxygen set as input variables, the acquired plankton distribution density with the same timestamp as output variables, and the relationship between the input variables and the output variables is fitted through a least square method to obtain a regression coefficient matrix beta, wherein the least square method is implemented by solving the formula To determine regression coefficients, X represents the input variable matrix, y represents the output variable vector, Representing covariance matrix, matrix Each element in the (a) is the inner product of the characteristic column corresponding to the input variable matrix and the target vector y; And (3) carrying out normalization processing on the newly acquired acoustic echo data, chlorophyll a concentration, oxygen demand data and dissolved oxygen data, and inputting the normalized acoustic echo data, the chlorophyll a concentration, the oxygen demand data and the dissolved oxygen data into a prediction model to obtain an estimation result of plankton distribution density.
7. The method for predicting the plankton density in the sea area based on the environmental factors and the acoustics of claim 1, further comprising a step S6 after the step S5, wherein the step S6 comprises the step of triggering an alarm if the estimation result of the plankton distribution density is higher than a preset density threshold value according to the estimation result of the plankton distribution density.

Description

Sea plankton density prediction method based on environmental factors and acoustics Technical Field The invention relates to the technical field of marine organism prediction, in particular to a marine plankton density prediction method based on environmental factors and acoustics. Background The health and stability of the marine ecosystem is critical to the sustainable development of the global environment and human society, while plankton, as the basis of the marine food chain, plays an irreplaceable role in maintaining ecological balance and continuity of fishery resources. However, in recent years, climate change, environmental pollution, and human activities such as marine facility construction have profound effects on the marine environment, resulting in significant changes in the distribution and density of plankton, which threatens not only the stability of the ecosystem but also may constitute a potential risk to the safety of the marine engineering. In the process of monitoring plankton density, the traditional method is often limited to measuring environmental factors such as chlorophyll a concentration, turbidity, dissolved oxygen and the like, and is difficult to cope with complex environmental changes in a large-scale sea area, and the requirement of real-time dynamic monitoring cannot be met. At present, acoustic wave reflection is used for monitoring plankton, but only the distribution situation of plankton groups can be reflected, but the signals of the plankton groups are influenced by various interference factors in the marine environment, so that the plankton groups are difficult to directly convert into reliable density data. Therefore, in the aspect of predicting the plankton density, the problem of inaccurate prediction caused by complex marine environment exists at present. Disclosure of Invention In order to overcome the defects in the prior art, the invention provides a sea plankton density prediction method based on environmental factors and acoustics, so as to solve the problems. The technical scheme adopted for solving the technical problems is that the sea plankton density prediction method based on environmental factors and acoustics comprises the following steps: s1, acquiring multi-source data from a target sea area, wherein the multi-source data comprises acoustic echo data and environmental factor data; s2, extracting noise level from the acoustic echo data; s3, judging whether the noise level exceeds a preset threshold value, and if so, denoising the acoustic echo data to obtain a pure acoustic signal feature set; s4, normalizing the characteristics and the environmental factor data in the pure acoustic wave signal characteristic set; s5, obtaining an estimation result of plankton distribution density through a multiple regression equation according to the normalized pure acoustic wave signal characteristics and the environment factor data. Preferably, in the step S2, the acoustic echo data of the same time stamp is classified by using a classification algorithm based on the environmental factor data to obtain the acoustic echo data of different tags; In the step S3, it is determined whether the noise levels of the acoustic echo data of the different tags exceed corresponding preset thresholds, respectively, where the preset thresholds are determined according to the categories of the acoustic echo data. Optionally, in the step S2, the environmental factor data includes chlorophyll a concentration, turbidity and dissolved oxygen, and in the step S2, the chlorophyll a concentration, turbidity and dissolved oxygen with the same time stamp are spliced to obtain spliced data, and the spliced data are combined to obtain an environmental factor set; Performing preliminary classification treatment on the environmental factor set through KMeans clustering of the Scikit-learn library to obtain a preliminary clustering result, generating corresponding pseudo tags according to the preliminary clustering result, and endowing each environmental factor data with corresponding pseudo tags to obtain environmental factor data with pseudo tags; the step of performing supervised classification using the trained XGBoost includes using the environmental factor data as input to the trained XGBoost, and outputting as classified environmental factor data and corresponding labels. Specifically, in the step S2, acoustic echo data of the same time stamp is acquired based on the classified environmental factor data, and a tag of the same class as the classified environmental factor data is given. It should be noted that, in the step S2, the acoustic echo data of different tags are combined into different acoustic wave feature sets, and the noise level is determined by calculating the standard deviation of the acoustic echo data in the different acoustic wave feature sets. Preferably, in the step S3, a preset threshold corresponding to the label of the acoustic wave feature set is obtained from a pr