Passive Millimeter-wave Image Restoration using Maximum Entropy Method with

Passive millimeter-wave (PMMW) image have an inherently poor resolution and highly blurred due to limited aperture dimension and the consequent diffraction limits thus requiring specialized sophisticated restoration algorithm before the image can be used for any useful purposes. In recent years many powerful algorithms based on Bayesian statistical concept are developed. These algorithms are nonlinear in nature and iteratively optimize the likelihood function to achieve high resolution. Some researchers also exploit the super-resolution capability of nonlinear algorithms. These algorithms are popular since simple digital implementation and robustness of performance to inaccurate estimation of sensor parameter. However, the convergence of iterations is slow and in practical implementation requires large number of iteration for getting required resolution. The quality of restoration and the extent of the achievable super – resolution of the image depends upon the how much and how accurate prior information about the true object is incorporated in algorithm. The prior information, which is used as constraint during restoration process, obtained by sensor characteristic, environmental condition at the time of recording, and true object related information. In this paper we discuss maximum entropy method with additional prior constraint flux conservation. Maximum entropy method is well known method and extensively used to for image restoration, but it has disadvantage that slow convergence and high computational requirement. Here we outline new approach using successive substitution to accelerate the convergence and reduce the computational load. A qualitative evaluation of this algorithm is performed with simulated data as well as actual radiometer image captured by 94 GHz mechanically scanned radiometer.