A Closely Spaced Objects Super-Resolution Method Using Array Camera Images

In the detection of spaced objects, the detected objects are usually far from the imaging system, resulting the weak radiation intensity and limited imaging area. Specially, the aliased projection of closely spaced objects further reduces the resolution of imaging results and restricts the subsequent high-level tasks. To solve this problem, a newly imaging system with array cameras has attracted more and more attention, which can provide complementary information of objects from multiple perspectives. In this paper, we study the geometric relationship between closely spaced objects and array cameras, and propose a closely spaced objects super-resolution method based on sparse reconstruction of array camera images. Thanks to the prior assumption of the spaced objects' sparsity and the transfer constraints between views of array cameras, our method can well super-resolve the obtained image and effectively estimate the number and location of the closely spaced objects.