AN ONBOARD IMAGE RECOGNITION MODULE FOR MONITORING SYSTEMS: METHODS AND ALGORITHMS

Abstract

The methods and algorithms of a compact on-board image recognition module to be used as a part of monitoring systems is presented. The main goal of the module is significantly reduction in the amount of image and video data, which should be stored on board and/or transmitted to a surface station. This is achieved by way of real-time recognition of frames of input video sequence and pre-selection of those similar to the given target image classes. It is supposed the module can be used as a part of remote monitoring systems based on drones and small-size satellites.

Keywords

• распознавание изображений
• мониторинг
• автоматическая классификация