A system mounted on a powered industrial vehicle for detecting classifying and tracking in real time at least one obstruction in the scene around the vehicle. The vehicle is capable of moving in multiple directions. The system includes a multiple cameras mounted on the vehicle, wherein the viewing zones viewed respectively by the cameras preferably encompass 360.degree. horizontally around the vehicle. Each of the cameras is operatively attached to an image processor, which processes the image frames acquired respectively by the camera. When a pedestrian is present in the viewing zone viewed by one of the cameras, the image processor attached to the one camera identifies in at least one of the image frames at least a portion of an image of the detected pedestrian.

In an image processor for synthesizing images that have been taken by multiple cameras mounted on a vehicle, even if any movable part of the vehicle has changed its state, the synthesized image will have no unnatural part. A pixel synthesizer receives the images that have been captured by the cameras included in an imaging section and generates the synthesized image by reference to a mapping table. Receiving a signal indicative of the state of a movable part such as the door or hood of the vehicle, a display mode changer disables at least one of the cameras that is taking an image to be affected by any change in the state of the movable part. For that purpose, the changer rewrites mapping data stored on the mapping table and thereby eliminates the image in question from the synthesized image.

A driving assistance apparatus includes: cameras; a plurality of virtual observing point converting unit; an image synthesizing unit for synthesizing image information supplied from this virtual observing point converting unit; an obstacle sensing unit such as an ultrasonic-wave sensor and a millimeter-wave radar; a safety area predicting unit; and a safety area superposing unit. Then, the safety area predicting unit predicts such a safety area where an obstacle is not present based upon both a distance "d1" and an approximate direction, which are obtained from the obstacle sensing unit. Also, the safety area superposing unit superposes the safety area predicted by the safety area predicting unit on an image synthesized by way of cameras to display the synthesized image superposed with the safety area.

A parameter selection section selects one from a plurality of image synthesis parameters stored in a parameter storage section according to an output of a vehicle state detection section. An image synthesis section reads out camera images from frame memories according to the selected image synthesis parameter and generates a synthesized mage showing a state of vehicle surroundings. A moving object detection section detects, in the thus generated synthetic image, a moving object region estimated as a region where a moving object is present.

A vehicle vision system that identifies pedestrians located proximate a vehicle. The system comprises a sensor array that produces imagery that is processed to generate disparity images. A depth map can be produced from the disparity images. Either the depth map or the disparity images are processed and compared to pedestrian templates. A pedestrian list is produced from the results of that comparison. The pedestrian list is processed to eliminate false pedestrians as determined by inverse eccentricity. The pedestrians are then tracked, and if a collision is possible, that information is provided to the driver and/or to an automated collision avoidance or damage or injury mitigation system.