Identifying Black Holes Through Space Telescopes and Deep Learning

The EHT has captured a series of images of black holes, demonstrating the possibility of detecting them through a radio interferometer. These images could provide valuable information about the gravitational environment near the event horizon. However, accurate detection and parameter estimation for candidate black holes are necessary. This paper explores the potential for identifying black holes in the ultraviolet band using space telescopes. Firstly, a data pipeline is established for generating simulated observations. Next, we present an ensemble neural network model for detecting and estimating the parameters of black holes with different angular sizes. The model can achieve a mean average precision value [0.5] of 0.9176 when reaching the imaging FWHM (θc) of the telescope and maintains its detection ability until 0.33θc. These results indicate that our methodology enables super-resolution recognition. The model accurately estimates the size of the accretion disk, the inclination angle, the positional angle, and the temperature. This provides a specific scheme for automatically detecting captured images using neural networks. Moreover, the validity of the model is assessed by analyzing the shadow of M87*, which indicates that the model can discriminate the black hole from background noise and other celestial objects with a confidence level of 0.639. Our work demonstrates the feasibility of detecting black holes in the UV band and provides a new method for the accurate and real-time detection of candidate black holes and further parameter estimation.