SAFT-based time-domain detection imaging method for internal defects of monocrystalline silicon

The invention discloses an SAFT-based time-domain detection imaging method for internal defects of monocrystalline silicon. The method comprises the following steps: allowing a phased-array transducer to move along a scanning surface and acquiring corresponding disperse echo signals of each data acquisition point via a probe; carrying out echo overlapping and summation on a certain point, regarding the processed point as a unit, and processing the rest points in a two-dimensional or three-dimensional surface one by one; when a sample point is located in a plane, mapping the sample point to a two-dimensional Cartesian grid, and when a sample point is located in a three-dimensional space, mapping the sample point to a three-dimensional Cartesian grid so as to determine a suspected defect area; and describing defect characteristics of the suspected defect area by using a 6dB-drop method. The method provided by the invention can clearly discriminate two barely separated holes with diameters of 0.8 mm and discovers the size and orientation quantification of a natural defect extremely approach the actually measured values of a cut sample. The method effectively improves spatial resolution and measurement precision of defects.