相位解包在光学测量领域有着广泛的应用，其中路径相关类解包方法通常计算速度较快，但抗噪声能力较弱；而路径无关类方法的鲁棒性好，但大多存在迭代收敛慢、计算量大的问题。因此，将区域重构技术引入解包运算中，得到一种新的路径无关类解包方法。该方法对于矩形区域数据能够直接求解解包相位，无需迭代过程；设计了采样-重组加速方法使运算量大幅减小，计算速度优于同类解包方法；使用Gerchberg式迭代过程能够解决不规则孔径内的数据解包问题，可在少量迭代后收敛。通过仿真验证了该方法在矩形区域内的直接求解精度和加速法的效能，通过对实验数据的解包验证了算法处理不规则孔径数据的能力。

The phase unwrapping methods are widely used in the field of optical testing, which can be divided into the path dependant type and the path independent type. Generally, the former is faster but more sensitive to the noises, while the latter is more robust but suffers from the problems of slow iterative speed and high time consuming. To solve this problem, a new path independent phase unwrapping method was proposed, which was based on the zonal reconstruction technique. Using this method the wrapped phase in the rectangle pupil can be unwrapped directly without any iteration, and the sample-recombining acceleration strategy was designed to reduce the calculation time greatly which made the algorithm faster than other path independent methods. In addition, to unwrap the phase in the irregular pupil the Gerchberg type iteration was used which can converge in a few steps. The precision of the direct unwrapping for the regular pupil was validated by the simulations, as well as the effect of the acceleration strategy. Besides, the ability to process irregular pupil problem was demonstrated by unwrapping the phase obtained in the experiment.