Research on Underwater Target Recognition Technology Based on Neural Network
نویسندگان
چکیده
At present, the underwater environment required by seafood aquaculture industry is very bad, and fishing operation completed artificially. In this environment, use of machine instead artificial development trend in future. By comparing characteristics different algorithms, multiscale Retinex algorithm (autoMSRCR) selected to deal with image color skew, blur, atomization, other problems. Labelimg software used annotate targets make data sets. Of these, 20% are as test sets, 70% training 10% verification The target detection network You Only Look Once Version4 (YOLOv4) based on convolutional neural networks (CNN) adopted paper. main feature extraction adopts CSPDarknet53 structure, fusion SSP, PANet carries out sampling convolution operations. prediction output extracted features carried through YoloHead network. After recognition model effect obtained testing identification accuracy sea cucumber urchin 90.8% 87.76%, respectively. Experiments show that can accurately identify specified organisms environment.
منابع مشابه
Research on Underwater Target Signal Detection and Recognition Processing Algorithm
Practical application of underwater target echo signal usually get disturbed a Gaussian noise and non-Gaussian noise, in view of the signal recognition problem, this paper proposes a double spectrum analysis based on wavelet transform domain method of weak signal and D-S data fusion algorithm. Through the study of double spectrum of the wavelet transform domain analysis to the signal processing...
متن کاملSea Clutter Target Recognition Based on Modular Neural Network
Sea clutter is a highly nonlinear signal; it has a non-stationary nature both in time and space. Many scholars found that sea clutter data are chaotic. This indicates that the system has some internal rules, but it is difficult to make an analytic formula. The BP neural network with self-organizing fuzzy spatial mapping ability of artificial happens to be a powerful tool for such kind of proble...
متن کاملRecognition of convolutional neural network based on CUDA Technology
For the problem whether Graphic Processing Unit(GPU),the stream processor with high performance of floating— point computing is applicable to neural networks, this paper proposes the parallel recognition algorithm of Convolutional Neural Networks(CNNs).It adopts Compute Unified Device Architecture(CUDA)technology, definite the parallel data structures, and describes the mapping mechanism for co...
متن کاملResearch on non cooperative target recognition technology based on Information Fusion Technology
In order to study non cooperative IFF to technology, this paper detailed illustrates the D-S theory in a multi-sensor data fusion principle and target recognition method, calculation method and practical model is established. By applying the theory of evidence (D-S method), we can solve the problem of data fusion in multi-sensor conditions. Aiming at the problem of multi-sensor target recogniti...
متن کاملPattern Recognition in Control Chart Using Neural Network based on a New Statistical Feature
Today for the expedition of the identification and timely correction of process deviations, it is necessary to use advanced techniques to minimize the costs of production of defective products. In this way control charts as one of the important tools for the statistical process control in combination with modern tools such as artificial neural networks have been used. The artificial neural netw...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Wireless Communications and Mobile Computing
سال: 2022
ISSN: ['1530-8669', '1530-8677']
DOI: https://doi.org/10.1155/2022/4197178