Abstract: Compressed Sensing is one of the most popular research fronts in recent years. This theory
has great influence in the fields of computational science, signal processing, and electronic information.
This technology has broad application prospects. Compressed Sensing breaks the “Nyquist Sampling
Law” in the field of signal processing. This method implements compression in sampling, achieving
the same effect as full sampling with few sample points. This technology puts forward new solutions
for the storage and transmission of massive information resources in the Big Data era, which is of great
significance. This paper introduces the theoretical basis and mathematical model of Compressed Sensing,
and attempts to apply compressed sensing technology to the construction of electronic resources in
digital libraries for the first time. This method of Compressed Sensing is based on Orthogonal Matching Pursuit algorithm, and can be used in the collection of scanned text resources and image, video electronic
resources. Compressed Sensing greatly reduces storage space and computation, and the compression
ratio is as high as 99.5%. It can be foreseen that the combination of Compressed Sensing and Artificial
Intelligence technology will become a subversive weapon in the construction of informational library
resources.

Key words: Compressed Sensing, Sparse sampling, Reconstruction algorithm, Intelligent library