End-to-end plankton database construction system and method

Abstract

The invention provides an end-to-end plankton database construction system and method. The system comprises a video collection processing unit, a target detection unit, a target recognition unit, anda plankton database construction unit, and the video collection processing unit records the video information of a plankton survival state and splits the video information into frame images; the target detection unit performs definition-based scanning and image segmentation to obtain a clear plankton region of interest; the target recognition unit is used for recognizing a target object in the plankton region of interest; and the plankton database construction unit fuses the image information of the same target object in different plankton interested domains to form a clear morphological structure feature database of each target object. According to the method, the data redundancy phenomenon can be solved, the memory consumption is reduced, plankton real-time monitoring is facilitated. Meanwhile, complete plankton morphological structure characteristics are provided, the accuracy is improved, and the method has important significance in biodiversity and marine ecological research.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to an end-to-end plankton database construction system and method. Background technique [0002] The development and application of underwater in-situ video acquisition technology has injected new vitality into the effective monitoring of the underwater environment, but it has also brought some new problems. On the one hand, the high-speed video acquisition system produces a large amount of redundant data, occupies a large amount of computer memory, reduces the efficiency of plankton research, and thus hinders the real-time study of plankton. On the other hand, the development of machine learning requires a large amount of training data, and manual calibration of massive data is time-consuming and laborious, and automatic processing methods need to be proposed urgently. At the same time, due to the influence of factors such as depth of field and target movement, the data o...

AI Technical Summary

Problems solved by technology

On the one hand, the high-speed video acquisition system produces a large amount of redundant data, occupies a large amount of computer memory, reduces the efficiency of plankton research, and thus hinders the real-time study of plankton

On the other hand, the development of machine learning requires a large amount of training data, and manual calibration of massive data is time-consuming and laborious, and automatic processing methods need to be proposed urgently

At the same time, due to the influence of depth of field, target movement and other factors, the data obtained in the high-speed microscopic video acquisition equipment often has partial blurring, which seriously affects the quality of the plankton database and thus affects the research of plankton by scientists.

Images