Within this bachelor's thesis, an algorithm for localizing the calibration object using data obtained from a 3D vision system has been elaborated in detail. The data collection process is achieved by using an established procedure that involves the collaboration of the Fanuc M710 robot and an integrated system (camera-laser). In the calibration process, the robot with a vision system is positioned at a series of known positions relative to the calibration board. The essence of the algorithm lies in the precise identification of coordinates corresponding to the edges belonging to the calibration board, which are used for calibrating the laser-camera system to map the positions of the laser detected by the camera into 3D spatial points. Before using these coordinates for system calibration, the RANSAC algorithm is applied to detect and eliminate potential errors and inaccuracies in the dataset, thereby improving the calibration accuracy. After applying the RANSAC algorithm, data processing is performed using a pre-developed neural network. The neural network's task is to test the success of the calibration, evaluating how accurately the system locates the calibration object in the real world. Calibration accuracy is crucial for the reliability and precision of all subsequent measurements and tasks performed by the system. Accurate calibration enables robots and image processing systems to precisely understand their environment and execute tasks with minimal or no errors.