Green tide area remote sensing monitoring method based on machine learning

• Main Authors: LIU FENG, FU XINGHU, SUN LIN, LU XIN, ZHAO JI, BI WEIHONG, FU GUANGWEI
• Format: Patent
• Language: Chinese; English
• Published: 30.08.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

The invention discloses a green tide area remote sensing monitoring method based on machine learning. The green tide area remote sensing monitoring method comprises the steps that an unmanned aerial vehicle carrying a hyperspectral imager is used for aerial photography in an experimental sea area; carrying out preprocessing work of image cutting, reflectivity correction, smooth noise reduction and spectrum enhancement on the aerial image; based on the spectral characteristics, carrying out band selection to extract spectral information and reduce the classification calculation amount; performing classification prediction on the aerial image by using a voting classifier, and calculating a green tide coverage area; according to the method, the voting classifier model taking the random forest, the support vector machine and the K-nearest neighbor method as three inputs is constructed based on the spectral feature information acquired by the field aerial hyperspectral image, and the prediction application of the