Deep-neural-network-based magnetic-field acquisition method for quantum sensing with nitrogen-vacancy centers

We propose a deep-neural-network-based method for acquiring the magnetic field using the nitrogen-vacancy centers in diamond. The nitrogen-vacancy ensembles have been recognized as a promising quantum sensor for detecting microscale magnetic fields, and the sensed magnetic field vectors can be obtained according to the resonance frequencies of the nitrogen-vacancy centers along four crystallographic orientations in the diamond. The proposed method is used to realize the conversion from the resonance frequencies to magnetic field vectors in sensor's coordinate system. A subnanotesla accuracy can be obtained in simulation data set for measuring the magnetic field strength, which is equivalent to geomagnetic field. The proposed method can have a considerable advantage in acquiring high-precision magnetic field vectors comparing to existing solutions as well as the analytical solution. The method is a potential technical supplement for nitrogen-vacancy-center-based magnetometry, and can push forward the development for nitrogen-vacancy magnetometer in applications that require large dynamic range and high accuracy.