The recovery of 3D volume of density from gravity field data is a key feature of geophysical and geological interpretations. Gravity field data inversion involves solving an underdetermined problem; therefore, large-scale data inversion is costly in time and memory consumption. Calculation efficiency is a primary concern for gravity field data inversion; therefore, multiple methods are considered and applied to increase the inversion efficiency. The solution for an inversion problem was formulated by incorporating constraints to obtain stable inversion results, and a new projected Barzilai-Borwein iterative algorithm was applied to accelerate convergence of the inversion method. To test the potential application of the projected Barzilai-Borwein iterative method, synthetic gravity data simulations and real data applications were carried out. Numerical performances and practical application indicate the fast convergence of projected Barzilai-Borwein iterative method and the increase in the calculated efficiency.

从重力场数据中恢复3D密度体积是地球物理和地质解释的关键特征。重力场数据反演涉及解决一个不确定的问题。因此，大规模的数据反转在时间和内存消耗上是昂贵的。计算效率是重力场数据反演的主要考虑因素。因此，考虑并采用了多种方法来提高反演效率。通过结合约束条件来获得反演问题的解决方案，以获得稳定的反演结果，并采用一种新的投影Barzilai-Borwein迭代算法来加速反演方法的收敛。为了测试预计的Barzilai-Borwein迭代方法的潜在应用，进行了合成重力数据模拟和实际数据应用。