Blue-green algae bloom prediction system and prediction method based on ACONV-LSTM and New-GANs combination

Abstract

The invention discloses a cyanobacterial bloom prediction system and prediction method based on the combination of ACONV-LSTM and New-GANs. The system comprises a cyanobacterial bloom chlorophyll a concentration level setting unit (200), an abnormal value removal processing unit (300), a continuous sampling time sequence image complementation unit (400), an Aconv-LSTM model (500) and a New-GANs model (600). According to the system, the chlorophyll a concentration grade is firstly matched to each remote sensing image to achieve unified data scale processing, and then linear interpolation is used for filling the deficiency to obtain a chlorophyll a concentration image with a complete time sequence; and inputting the complete chlorophyll a concentration image into an attention convolution long-short term memory network and an improved generative adversarial network, thereby predicting chlorophyll a concentration outbreak in future time.

Description

technical field [0001] The invention relates to cyanobacteria bloom prediction and prediction in the technical field of water quality monitoring, more particularly, to a cyanobacterial bloom prediction system and prediction method based on the combination of ACONV-LSTM and New-GANs. Background technique [0002] With the development of urbanization and industrialization, urban sewage, industrial sewage and other water sources containing nutrient-rich substances enter the lake, resulting in aggravating the nutrient load of the water body. In the ecosystem, algae become the dominant species and multiply, destroying the ecological balance of the water body and eventually leading to the deterioration of water quality. In summer in my country, there are many lakes and rivers that have different degrees of cyanobacterial blooms. These cyanobacterial blooms cover the water surface or are cleaned up on the shore. However, in a high temperature environment, it is easy to form a stenc...

AI Technical Summary

Problems solved by technology

[0006] In order to solve the technical problems such as time-consuming and laborious sampling of chlorophyll a data for prediction of cyanobacteria bloom prediction, small sampling data coverage and small prediction range, the present invention adopts attention convolution long-short-term memory network (ACONV-LSTM) and improved generative formula Combined with the confrontation network (NEW-GANs), the time series prediction of the remote sensing inversion image of the chlorophyll a concentration is performed, and the chlorophyll a concentration is predicted in the future by inputting the inversion image of the historical chlorophyll a concentration remote sensing image. The present invention is based on ACONV-LSTM and The cyanobacterial bloom prediction system combined with New-GANs can realize the prediction of cyanobacterial blooms in the whole water area, and the required chlorophyll a data does not need to be sampled on site, providing a new technology for algal bloom prediction

Images