Red tide biological picture automatic identification device and identification method

Abstract

The invention relates to a red tide biological image automatic recognition device and method. A light emitting diode, a light source lens, a flow chamber, an objective lens, a depth-of-field lens and a camera are provided; the light-emitting diode is located at the focal point of the light source lens, the flow chamber is located between the light source lens and the objective lens, and the depth-of-field lens is located between the objective lens and the camera. The recognition method is as follows: image acquisition, preprocessing, target detection, image segmentation, image feature acquisition, target classification and recognition, and result output. Through the processing of the lens, a uniform background light source can be generated, which is conducive to obtaining a good image. The far-focus objective lens is used to increase the operating space between the flow chamber and the objective lens, giving the flow chamber a wider development space and improving the application range of the instrument. It can automatically analyze the image, improve the automation of the instrument, and also meet the requirements of red tide biological on-site monitoring. It is expected to realize real-time monitoring of red tide production and elimination process, and provide early warning of red tide danger on this basis.

Description

technical field [0001] The invention relates to an image recognition device and a recognition method, especially to distinguish red tide organisms from plankton groups in water bodies by using image features, and then identify them according to their characteristics, so as to realize the identification of red tide organism species, and Due to the particularity of the source of the acquired image, the red tide biological image automatic identification device and its identification method can realize on-site monitoring of red tide organisms. Background technique [0002] Red tide is one of the common and important disasters in my country's coastal waters, and it is an important project for marine ecological environment monitoring. Since the 1970s, the frequency of red tides in my country has increased by three times every ten years. In 2002, there were 79 red tide events, with a cumulative area of ​​more than 10,000 square kilometers, causing serious ecological, resource, envi...

AI Technical Summary

Problems solved by technology

However, due to the lack of red tide biological database and image recognition system, it is impossible to automatically analyze the acquired images, and it is still impossible to automatically identify and monitor red tide organisms

Images