合稀疏贝叶斯学习(SBL)和可压缩传感理论(CS)，给出一种在噪声测量条件下重建可压缩图像的方法。该方法将cS理论中图像重建过程看作一个线性回归问题，而待重建的图像是该回归模型巾的未知权值参数；利用sBL方法对权值赋予确定的先验条件概率分布用以限制模型的复杂度，并引入超参数-

Hop sparse Bayesian learning ( SBL ) and compressible sensing theory ( CS ) , give a compressible image reconstruction in the noise measurement conditions . The method of the CS theory image reconstruction process as a linear regression problem , the image to be reconstructed is unknown weighting parameters of the regression model towel　 SBL method to determine the weights given a priori probability distribution to limit the complexity of the model and the introduction of the hyper-parameters