K-space reconstruction method and magnetic resonance imaging method

The invention provides a K-space reconstruction method. The K-space reconstruction method includes: parallelly collecting K-space data including data should be collected and calibration data; extracting the calibration data to form a first data set, and executing pixel merging on the first data set to acquire a second data set, wherein the first data set and the second data set are identical in scanning field of view; utilizing the second data set to compute a coil merging coefficient, and filling the under-sampling K-space based on the coil merging coefficient so as to acquire a full-sampling K-space. By the method, signal to noise ratio of data is increased, computational accuracy of the coil merging coefficient is optimized, and image artifacts caused by fitting error are reduced.