Foresight Autonomous Holdings Ltd. announced that the Company has developed significant advanced features for its four-camera QuadSight vision system. The new features were developed to meet customer requirements following successful evaluation of several QuadSight system prototypes purchased over the past year. These capabilities will significantly improve the QuadSight system’s performance in accurately measuring obstacle size, location and distance, even in harsh lighting and weather conditions. The QuadSight system’s new features include automatic calibration, 3D point cloud and multispectral sensor fusion. Automatic calibration: The main challenge using stereo vision technology is the ability to create a precise 3D depth map in a dynamic environment, achieved through synchronization and calibration of the stereoscopic cameras. In order to obtain precise measurements of the vehicle’s surroundings, it is necessary to estimate the rotation of one camera compared to the other for each frame, a calculation known as relative pose estimation. Foresight has developed a proprietary solution known as automatic calibration which uses unique algorithms to successfully assess relative pose estimations. Automatic calibration is fundamental for creating an accurate stereoscopic 3D perception while ensuring system robustness. Point cloud is a set of data points which generates accurate pixel location and distance for complete 3D mapping. To the best of the Company’s knowledge, QuadSight is the only sensor which generates an infrared-based point cloud. Foresight’s technology generates a high-resolution depth map which is converted to a high-resolution 3D point cloud; this provides accurate information on the vehicle’s surroundings including the location and distance of any object in the field of view. In addition, the Company intends to further develop its unique point cloud solution to fuse data from both visible light and longwave infrared channels. Point cloud provides raw 3D data that can be used for obstacle detection, terrain analysis and fusion with other autonomous vehicle sensors. Multispectral sensor fusion: The QuadSight system combines two channels which consist of stereo visible-light cameras and stereo long-wave infrared thermal cameras to provide accurate obstacle detection in harsh lighting and weather conditions. The system performs fusion between the two stereoscopic channels in order to generate an accurate depth map while reducing false alerts. In addition, the system can decide which channel to choose, in real time, according to the relevant scenario.