Detekce lodí ve video sekvencích

Abstract

This work investigates methods for analyzing video sequences from cameras at shipping lanes and wharves with the aim of detecting and controlling ships passing through them. Based on real data from cameras located in the harbor and along the shipping route, these methods are systematically evaluated and compared, identifying their advantages and limitations. The results of this analysis are then used to design an optimal solution for ship detection. The proposed functional solution thus uses modern image analysis technologies and enables efficient detection of ships. The resulting application also allows the user to choose and use prepared object detection methods on selected camera data.