快速射电暴（Fast Radio Burst，FRB）探测的实时性及位置估计的精度是触发多波段观测、联合干涉测量以实现FRB宿主起源证认和物理本质认识的关键。视场及定位精度已成为FAST快速探测及精确估计FRB位置的瓶颈，亟待解决。.针对FRB探测的前沿需求，课题研究基于焦面场特征深度理解的FAST探测FRB理论与技术。以焦面场为研究对象，揭示其几何、纹理特征在时域、频域及极化域的特征，建立射电目标的位置与焦面场分布特征的映射关系；针对FRB持续时间短、频率成分丰富的特点，建立其对应的FAST焦面场在欠/过采样条件下特征提取对望远镜探测FRB的视场及定位精度的作用模型；开展FRB在焦面场映射特征的机器理解研究，结合压缩感知、深度学习算法，建立焦面场信息增强及高精度定位方法体系并完成缩比实验，突破定位精度受限于波束宽度的瓶颈。课题成果也适用于FAST对其它瞬态目标的探索与定位。

The real-time requirement and the precision of the position estimation of the Fast Radio Burst (FRB) detection are the keys to triggering multi-band observations and joint interferometry to realize the FRB host origin identification and physical essence understanding. The field of view and positioning accuracy has become the bottleneck of rapid detection and accurate estimation of FRB's position with Five-hundred-meter Aperture Spherical radio Telescope (FAST), and it is urgent to be solved..For the frontier demand of FRB detection, the research focus on the detection FRB theory and technology using FAST with the deep understanding of the focal plane field feature. Taking the focal plane field as the discussion object, the characteristics of the geometric and texture features in the time domain, frequency domain, and polarization domain are revealed, then the mapping relationship between the position of the radio target and the distribution characteristics of the focal plane field is established. For the transient duration and rich frequency components features of the FRB, a model of the corresponding FAST focal plane field under the down/up sampling conditions feature extraction to the telescope detection field of view and positioning accuracy will be established. With the machine understanding of FRB’s mapping characteristic in the focal plane field, combined with Compression Samping, Deep Learning, focal plane information enhancement and high-precision positioning method system will also be proposed. The related experiments will be carried out to verify the all the proposed theory and method, which will break the bottleneck that the current positioning accuracy limited by beamwidth. The results will also be ready to apply to the exploration and positioning of other transient targets with FAST.