A neural network-based method for detection and tracking statistics of near-shore seabed fish

Abstract

The invention discloses a neural network-based nearshore seabed fish detection and tracking statistical method. The invention generates clear images by performing FcycleGAN image migration processing on the input underwater real-time video, and then inputs it into the basic neural network Darknet53 for processing to generate clear images. The features of the fish in the video are extracted, mainly including the shape features and texture features of the fish. The detection branch is detected in two stages, and finally the specific position and species of the fish are output. The tracking branch outputs the Jacobian matrix and distance vector of the swimming fish particles. Then, a certain range of fish in the predicted position is matched with the fish in the previous position, so as to obtain the position, category and number of the fish in each picture.

Description

technical field [0001] The invention relates to the field of seabed exploration and detection, in particular to a neural network-based nearshore seabed fish detection and tracking statistical method. Background technique [0002] The ocean is rich in biological resources; therefore, coastal countries are vigorously developing marine pastures, especially fishery-enhanced marine pastures. The Food and Agriculture Organization of the United Nations, the Food and Agriculture Organization of the United Nations, recorded 28.7 million tonnes (US$67.4 billion) of global food fish production in marine pastures in 2016, accounting for 49.5% of total world aquaculture production in 2016. Currently, offshore fishing is being Over-exploitation, the aquaculture industry is also in a state of saturation; therefore, marine ranching management is considered to be an important way to solve the decline of fishery resources, however, there are also some problems (such as overfishing, ecosystem ...

AI Technical Summary

Problems solved by technology

Finally, a hybrid neural network model was constructed, however, the main limitation of this approach is that it treats fish as particles and cannot classify fish types, Marini et al. (2018) developed a genetic programming-based content-based image analysis methods, however, when a large number of fish gather in front of the camera, the crowded scene limits the recognition efficiency, and when these gatherings are particularly dense, individuals often overlap each other, which increases the false negative rate

The current existing technology is aimed at the classification and recognition of underwater fish images. The problem with them is that this type of method is not conducive to the statistics of fish schools, and will cause serious errors. To distinguish whether it is the same fish as the previously identified fish, so there will be redundant errors when counting the number

Images