Models and algorithms for a real-time hybrid image enhancement methodology

摘要

The paper presents models and algorithms for a hybrid method of image enhancement. The proposed framework involves processing of the digitized image to transform the input signal from image sensors such that the digitization of the transformed analog signal yields the enhanced image. First, an analytical framework which yields robust algorithms for image enhancement (following our approach) is established. Unlike many conventional methods, the approach allows the user to specify certain statistical measures for desired gray values in the enhanced image. An appealing feature of the proposed method is its ease of implementation using off-the-shelf hardware such as Analog-to-Digital Converters. To support our models and analyses, we perform character recognition on a set of industrial parts.