One of the important topics in seismic exploration is seismic velocity analysis. Increasing the accuracy of the best velocity model in prestack data has been studied by various methods such as velocity spectrum analysis, interactive velocity analysis, horizon velocity analysis, and wavelet transform. In this research, increasing the accuracy of speed spectrum peaks selection based on three types of compression algorithms including circular averaging filter, motion blur image and averaging filter has been investigated. In the artificial data, the desired result was not achieved, but in the real data, circular averaging filter algorithm has achieved a favorable compression compared to motion blur and averaging algorithms.