Informational and extreme machine learning for onboard recognition system of ground objects

The method of machine learning of the onboard system of an unmanned aeronautical complex for autonomous identification of frames formed by an optical channel of digital image of the region is considered. The task of informational synthesis of the on-board frame identification system is being solved within the framework of information-extreme intellectual data analysis technology, which is based on maximizing the information capacity of the system in the process of machine learning. According to the categorical model proposed in the functional approach to the modeling of cognitive processes, an algorithm of information-extreme machine learning was developed and programmatically implemented. To ensure the invariance of the decisive rules for the displacement and rotation of objects within the frames of the formation of the input mathematical description was carried out by processing the frames of the digital image of the region in a polar coordinate system. The modified Kullback criterion was used as the criterion for optimization of machine learning parameters. The implementation of the proposed machine learning algorithm was based on the identification of digital frames of the region, which characterized various natural and infrastructural objects, among which roads were considered as a zone of interest. According to the results of the frame identification analysis, the full probability of correct classification is close to one.