Environment perception by fusing laser range sensor and stereo camera data

Senzorski sustavi koje čine kombinacija laserskih senzora i kamera dobiva na popularnosti kao izbor za percepciju okoline na mobilnim robotima. Kako bi se ostvarila uspješna i pouzdana percepcija okoline potrebno je povezati informacije iz oba senzora. Također je potrebno kalibrirati senzore, odnosno pronaći njihov odnos u prostoru. Zadatak rada je dobiti informacije o udaljenosti prepreka za oba senzora te odrediti njihov odnos u svrhu fuzije podataka. U ovom radu je prvo dan uvod u Robotski operacijski sustav (ROS). Zatim slijedi objašnjenje metode korištene za kalibraciju kamere i objašnjenje parametara dobivenih kalibracijom. Na kraju je opisana metoda kalibracije lidara i stereo kamere te su prikazani rezultati dobiveni fuzijom podataka iz oba senzora.Sensor system that consists of a combination of laser sensors and cameras has gained in popularity as a choice for environment perception for mobile robots. To ensure a successful and reliable environment perception it is necessary to fuse information from both sensors. Also it is required to perform sensor calibration, in other words find their spatial relation. The goal of this thesis is to acquire information about the distance of obstacles from both sensors and to determine their relation for the purpose of data fusion. In this thesis, firstly you are given an introduction to Robot Operating System (ROS). Then follows the explanation of the method used for camera calibration and the explanation of parameters acquired from the calibration. Finally, the method used for calibrating lidar and stereo camera, is explained and results obtained from data fusion are presented