Processing of Sonar Image Based on Compressive Sensing

Xu, Zhi Jing; Dai, Huan; Cao, Pei

doi:10.4028/www.scientific.net/amr.301-303.719

Cited by 2 publications

(1 citation statement)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In sonar image compression processing, reference [1] put forward the image data compression method of sonar image by performing the Discrete Cosine Transform (DCT) and subsection quantization, the object of this method is image data acquired by real-time imaging sonar, but DCT transform is easy to have block effect; Reference [2] put forward the image compression method which implemented coding of the EMD with DCT of the non-uniform sampled blocks and mode-image coding, but the reconstruction of image existed block effect even implementing the non-uniform sampling, particularly in small size images, low-frequency mode and high threshold condition; Reference [3] adopted compressive sensing theory based on the characteristics of sonar image, wavelet transform was used to sparse decompose the image, the measurement and reconstruction used conventional algorithm, it proves that the quality of the reconstruction image and PSNR have improved compared to the traditional compression based on the wavelet transform, but the edge and contour of reconstruction images based on wavelet are fuzzy. Affected by noise, sonar image has small overall average and low contrast, so the compression should try to keep the image edge and contour in the compression process.…”

Section: Introductionmentioning

confidence: 99%

Underwater acoustic image compressive sensing algorithm research based on Bandelets transform

Cai

Chen

Yuan

2016

OCEANS 2016 - Shanghai

View full text Add to dashboard Cite

Sonar detection has important application in underwater topography exploration and underwater scene detection, inevitably sonar image is transmitted through the underwater acoustic channel, but the particularity of underwater acoustic channel has brought a lot of limitations to underwater image transmission, so how to make effective collection, compression and transmission of underwater image to promise that the image processing do not affect the reconstruction results of sonar images is a problem worth studying. In this paper, the compressive sensing theory is used in the sonar image compression processing: Bandelets transform is used to sparse decompose the image, and select Gaussian random matrix as the observation matrix, use the Orthogonal Matching Pursuit (OMP) algorithm to reconstruct the image. The compressive sensing based on Bandeltes transform save the edges and contours information of the sonar image effectively, the image reconstruction result is also better than that of the image based on wavelet transform. Based on the low frequency characteristic of sonar image, this paper also puts forward the improved algorithm combined with the wavelet transform and compressive sensing, the algorithm first obtain low frequency information and high frequency information of sonar images by performing the wavelet transform, and then collecting low frequency information and high frequency information by different sampling rate finally reconstruct the low frequency image and high frequency image and recovery the original sonar image by performing the inverse wavelet transform. The experiments result show that the improved algorithm can achieve a better reconstruction result with lower comprehensive sampling rate. Compared with original compressive sensing method, the assessment parameters of image quality SSIM and PSNR has promoted. The improve algorithm reduces the amount of transmission data, improve the transmission efficiency, and can well reconstruct the original sonar image, also has high robustness, provides a convenient method for the sonar image transmission.

show abstract

Section: Introductionmentioning

confidence: 99%