Compressive X-ray Tomosynthesis based on Exposure Sequence Optimization

X-ray tomosynthesis technique can reconstruct three-dimensional object of interest from two-dimensional projection measurements. This paper proposes a compressive X-ray tomosynthesis imaging method to inspect the moving object on the conveyor. Coding masks are used to modulate the intensity distribution of X-ray sources and reduce the radiation dose. Low-cost uniform linear array detector is used to collect measurement data. Considering the rigid motion of the object, the proposed method uses a dynamic imaging model and a reconstruction framework based on the variation of the sensing geometry. Subsequently, an optimization method is proposed to design the exposure sequence of X-ray sources based on the genetic algorithm. Optimizing the exposure sequence can ameliorate the condition of the ill-posed inverse reconstruction problem. Simulations show that the optimized exposure sequence can effectively improve the reconstruction performance of the compressive X-ray tomosynthesis.