Rapid low-memory image compression perception method

Abstract

The invention discloses a rapid low-memory image compression perception method and belongs to the field of image processing. According to the invention, by use of a low-order random observation matrix, local sampling and blocking reconstruction are performed on an original signal, so the disadvantage that an observation matrix needs to occupy a lot of storage space and memory space in the traditional observation perception method is overcome, the disadvantage of poor timeliness of large-size image construction is overcome, storage space required by the random observation matrix is reduced, calculation complexity of the reconstruction is reduced and timeliness of the reconstruction is improved.

Description

technical field [0001] The invention belongs to the field of image processing, and in particular relates to a fast and low-storage image compression sensing method. Background technique [0002] Since the compressed sensing (Compressed Sensing, CS) theory was put forward, researchers at home and abroad are constantly improving the theory of compressed sensing and expanding the application of compressed sensing. There are always problems to be improved and solved in the research of sampling and reconstruction. [0003] First, in terms of non-correlated observation of signals, although the projection method of the random observation matrix has theoretically perfect characteristics, due to its random characteristics, the hardware implementation, storage allocation and reconstruction algorithm construction of the random observation matrix all require Occupies a large amount of storage space and memory space, which is greatly restricted in practical applications. To this end, r...

AI Technical Summary

Problems solved by technology

[0004] Second, in the optimal reconstruction process of the signal, as the size of the image increases, the computational load of the reconstruction algorithm will increase exponentially, making the reconstruction process of the entire image very time-consuming, and the reconstruction process The amount of data storage and memory usage will increase exponentially, greatly reducing the real-time performance of compressed sensing applications

For example, when the sampling r

Images