No abstract

The primary task of the 1.26 m telescope jointly operated by the National Astronomical Observatory and Guangzhou University is photometric observations of the g, r, and i bands. A data processing pipeline system was set up with mature software packages, such as IRAF, SExtractor, and SCAMP, to process approximately 5 GB of observational data automatically every day. However, the success ratio was significantly reduced when processing blurred images owing to telescope tracking error; this, in turn, significantly constrained the output of the telescope. We propose a robust automated photometric pipeline (RAPP) software that can correctly process blurred images. Two key techniques are presented in detail: blurred star enhancement and robust image matching. A series of tests proved that RAPP not only achieves a photometric success ratio and precision comparable to those of IRAF but also significantly reduces the data processing load and improves the efficiency.

The state space explosion restricts the error detection of concurrent software. The abstraction can provide a solution to avoid state space explosion, but it is easy to ignore important details, resulting in inaccurate detection results. This paper proposes a methodology of fine-coarse-grained automatic modelling for Java source programs. By the principle that the execution details of property-unchecked, non-interactive, and unrelated statements do not affect the model checking results, we model coarse-grained model fragments for such statements, while fine-grained model fragments for property-checked, interactive, and related statements. Our method reduces the model and state space and ensures the error detection of the source program based on model checking. Moreover, we prove the equivalence of the fine-grained model, the coarse-grained model, and the program. Finally, this paper gives an experiment to verify the effectiveness of the proposed method.

No abstract