Author(s)

Renying Song

Corresponding Author

Renying Song

Abstract

In order to analyze the types and quantities of harmful algae contained in seawater, early monitoring and forecasting of red tide algae development was carried out, and a computer automatic identification system for red tide algae images was developed. Image processing technology was used to extract algae image morphology and texture features, and genetic algorithm was used to select features. On this basis, a neural network is used to establish a classifier to classify and identify algae images. The results show that the system can effectively improve the learning ability and classification performance, and achieve good classification and recognition of the three main algae that cause red tide. The analysis results are less than the results of manual counting and recognition.

Keywords

Automatic Analysis, Red Tide Algae, Image Identify