To address the problem of how to select appropriate edge sampling points in the phase distribution map, in this paper, we analyze the advantages and disadvantages of different processing schemes and propose a sampling strategy based on phase imaging theory and experiments. Combined with the phase imaging mechanism, several gradient-based classical edge detection operators were used to analyze the source of uncertainty in the position of sampling feature points and its influence on the detection results. Actual samples showing a morphological structure with different complexities were selected to conduct experimental research. Furthermore, the influence of different sampling points on the sample size calculation results was compared and analyzed. Subsequently, the causes of various errors were investigated and their theoretical correlation and explanation were presented. Based on the results of a comparative analysis of multiple groups of experimental data, a morphological structure feature extraction scheme suitable for samples with different complexities was proposed. The processing efficiency of this scheme can meet the needs of a rapid real-time detection.