Tensor imaging can provide more comprehensive information about spatial physical properties, but it is a high-dimensional physical quantity that is difficult to observe directly. This paper proposes a fast-transform magnetic tensor imaging method based on the NV magnetic detection technique. The Euler deconvolution interprets the magnetic tensor data to obtain the target three-dimensional (3D) boundary information. Fast magnetic vector imaging was performed using optical detection of magnetic resonance (ODMR) to verify the method's feasibility. The complete tensor data was obtained based on the transformation of the vector magnetic imaging data, which was subsequently solved, and the contour information of the objective was restored. In addition, a fast magnetic moment judgment model and an angular transformation model of the observation space are developed in this paper to reduce the influence of the magnetic moment direction on the results and to help interpret the magnetic tensor data. Finally, the experiment realizes the localization, judgment of magnetic moment direction, and 3D boundary identification of a micron-sized tiny magnet with a spatial resolution of 10 µm, a model accuracy of 90.1%, and a magnetic moment direction error of 4.2°.