Jellyfish Recognition And Density Calculation Based On Image Processing And Deep Learning

In order to monitor the invasion of marine organism in real time, the combination of image processing and deep learning is used to realize species identification and density- calculation of marine life. Underwater camera collects real-time image data within the monitoring range, and realizes the end-to-end jellyfish recognition by using depth learning. First, the convolution neural network is designed and improved to obtain the convolution neural network composed of two convolution layers, two pooling layers and full connection layers. Then, the accuracy performance results and training model are obtained by training, and the test sample pictures are used for prediction. According to the characteristics of organism image in the non-uniform light field of turbid water, the image sharpening, edge detection, edge closure, hole tilling and other calculations are carried out on the gray image data, and the binary image separated from the target and background is obtained. The real-time measurement of the estimation of marine organism density is realized. The results show that the method proposed in this paper can effectively apply the calculation of marine organism density and the detection of marine organism species Measurement. The results of this study can provide the main reference for the early warning of organism invasion in the near sea water.