Design And Implementation Of Pulsar Recognition System Based On Deep Learning

Five hundred meters Aperture Spherical Radio Telescope, called FAST, is one of China's most important science and technology infrastructure. One goal of FAST is to discover 5000 pulsars and the candidate pulsar signal is about 1 million or more after preliminarily estimating. It is unrealistic to identify a small number of real pulsar signal in a extreme number of candidates pulsar signal rely on manpower because it will consume lots of human resources, financial resources and time. It will greatly improves the efficiency of the identification of pulsars by using computer, for it will greatly reduce the human and financial expenditure.Deep learning is a hot topic of machine learning in recent years and has achieved great success in the field of image recognition, such as handwriting recognition, face recognition, plate number recognition etc.This paper studies the pulsar signal recognition method based on deep learning and then completes the pulsar recognition software system.The main work of this paper includes:(1) Through referencing to human recognition algorithm, this paper research pulsar recognition algorithm base on deep learning algorithm. Experimental results showed that the pulsar identification algorithm that this paper proposed is feasible.(2) This paper has designed and implemented the pulsar recognition software system.