A distributed algorithm for adaptive regulation of image processing parameters

The authors describe an algorithm able to regulate and control the parameters that manage some common processing algorithms, to obtain the best data quality. The processing chain to be regulated consists of an acquisition module and two filtering processes. The system has been implemented as an expert system, which, in addition to the processing units, includes a regulation module. Each regulation module has been specialized with knowledge about the parameters regulating the related processing unit. Criteria to judge the quality of data at each level have been devised to develop methods to select and regulate the processing parameters to be adjusted. The system is described specifying the knowledge-based architecture and single units where knowledge is embedded. The functioning of each processing module is illustrated by defining the parameters used to judge data and those used to regulate the units. Results on a set of indoor images, acquired by using different camera parameters, are reported.<<ETX>>