Lens antenna array is considered as an effective beam selection mechanism in millimeter wave massive multiple input multiple output systems. Efficient channel estimation (CE) algorithms are required to use the advantage of the beam selection paradigm. Recently, compressive sensing-based algorithms are used to utilize existing sparsity for CE in these systems. Among them, orthogonal matching pursuit (OMP) and support detection (SD) are the most popular ones. These two popular algorithms have their own advantages and disadvantages. In this paper, we propose to use OMP and SD together for better CE. Simulations validate that the proposed algorithm enhances the CE quality over the conventional algorithms. These simulations are tested over two popularly used channel models.