Detection of Distant Obstacles Using Computer Vision on a Unmanned Surface Vehicle

Delo predstavlja pristop k razvoju metode, s katero bo brezpilotno plovilo detektiralo ovire, ki so zunaj dosega stereo para kamer. Algoritem temelji na geometriji dveh pogledov s konceptom strukture iz gibanja. Poleg kamer upošteva tudi izhode ostalih senzorjev na robotu, ki se uporabljajo za določanje lege kamere. Predlagan pristop se opira na iskanje primernih parov v podatkovnem zalogovniku, ki hrani pretekle podatke zajetih slik. Prostor iskanja parov slik zmanjša z uporabo rojenja k-povprečij in drugih geometrijskih lastnosti lege kamer. Izhod algoritma je naraščajoči oblak točk, ki je na voljo metodam detekcije ovir in planiranja poti. Opravljeni so bili eksperimenti, s katerimi sta se ocenila ponovljivost detektorjev in vpliv negotovosti detekcij na rekonstrukcijo oblakov točk.This thesis proposes an approach to a method, used by an Unmanned Surface Vehicle (USV) to detect obstacles, that are outside the range of onboard stereoscopic camera pair. Algorithm is based on a concept of Structure from motion using two-view geometry. It also utilizes data from other available sensors, such that are used to assess the camera pose. The proposed approach relies on finding suitable pairs in the data buffer, that contains various data of previously captured images. Search of image pairs is reduced by implementing k-means clustering with respect to other geometric constraints on camera poses. Algorithm is building a growing point cloud, which is available to obstacle detection and path planning procedures. We have conducted experiments, which evaluate the detector repeatability and determine the effect of detection uncertainty on point cloud reconstruction