System for automatic recognition, identification and measuring of floating vessels

Abstract

The subject of the invention is a system for automatic recognition, identification and measurement of the length of a vessel (1). The system basically consists of a stereo camera (3) with a field of view (2) of at least 120 degrees and in which the movement of the vessel (1) is monitored. At the moment when the vessel (1) arrives at a point perpendicular to the camera (3), the system creates disparity images based on which the vessel (1) is reconstructed in space by determining the distance of each pixel (z) at a certain coordinate (x,y). This results in a depth image and accurate information about each pixel in space. The machine learning model determines the so-called mask, i.e. the pixels that are the vessel and which are the environment. For the pixels that belong to the vessel (1), the positions (x,y,z) in space are determined, which measures the physical dimensions of the vessel (1). The subject of the invention is a system for automatic recognition, identification and measuring of the length of the vessel (1).The system comprises basically a stereo camera (3) which has a field of vision (2) of at least 120 degrees and in which the movement of the vessel (1) is monitored. At the moment of the vessel (1) coming to a point perpendicular to the camera (3), the system creates disparity images based on which the vessel (1) is reconstructed in space by determining the distance of each pixel (z) at a certain coordinate (x, y). This provides an in-depth image and accurate information about every pixel in the space. The machine learning model determines the so-called mask, i.e. the pixels that are the vessel and which are the environment. For the pixels belonging to the vessel (1), the positions (x, y, z) in space are determined, which measures the physical dimensions of the vessel (1).