Abstract :
We have designed and implemented a navigation and obstacle avoidance system for our camera-guided patrolling vehicle, which was installed with an 8-node multiprocessor for the image processing, digital communication and motor control. We use the middle line of the road as the guiding mark for the video navigation. In addition to the ultrasonic obstacle avoidance system, which fails in the range outside 5 m, we add the video device to get around obstacles on the road. We consider two cases according to the environmental light; one is the dark case, the other is the bright case. In the dark case, we install one laser and one CCD camera in the front edge of the vehicle. They are separated 60 cm apart and aim at the same point located 10 m in front of the vehicle. If the image of the reflected laser light on the obstacle appears in the middle of the screen, it means the obstacle is 10 m ahead. The vehicle then executes an avoidance action. In the bright case, we replace the laser with one more camera. If both the cameras show the image of the obstacle in the middle of the screen, the vehicle then executes the avoidance action. The performance of the navigation and obstacle avoidance is reported
Keywords :
CCD image sensors; active vision; computerised navigation; digital communication; image processing equipment; laser ranging; path planning; road vehicles; 0 to 5 m; 10 m; 60 cm; 8-node multiprocessor; CCD camera; avoidance action; bright case; camera-guided patrolling vehicle; dark case; digital communication; environmental light; guiding mark; image processing; motor control; performance; reflected laser light; road mid-line; ultrasonic obstacle avoidance system; video device; video navigation;
