Real-time implementation of a novel detail enhancement algorithm for thermal imager

Abstract

This paper presents real time implementation of a new detail enhancement algorithm to effectively visualize High Dynamic Range (HDR) InfraRed (IR) images on Low Dynamic Range (LDR) display with reduced noise amplification. Rendering real world IR images on state of the art display devices requires compression of HDR IR images without losing the perceptibility of details. The proposed algorithm is a combination of preprocessing module for detail enhancement and Dynamic Range Compression(DRC) for LDR display. In the preprocessing module, the base and detail components of input HDR image are obtained using bilateral filtering. The detail component of IR image is scaled and added back to its base component and then the resultant is compressed using DRC method for display on LDR monitors. MATLAB simulation and Real-Time implemented results show a significant improvement in the results obtained when a given DRC method is preceded by the proposed preprocessing module. The DRC algorithms referred here are AGC, classical Histogram Equalization, Histogram Projection and Adaptive Plateau Equalization. The real time implementation procedure of current proposed algorithm using xilinx Virtex-4 FPGA is explained.