Techniques for road scene primitive detection using a vehicle camera system are disclosed. In one example implementation, a computer-implemented method includes receiving, by a processing device having at least two parallel processing cores, at least one image from a camera associated with a vehicle on a road. The processing device generates a plurality of views from the at least one image that include a feature primitive. The feature primitive is indicative of a vehicle or other road scene entities of interest. Using each of the parallel processing cores, a set of primitives are identified from one or more of the plurality of views. The feature primitives are identified using one or more of machine learning and classic computer vision techniques. The processing device outputs, based on the plurality of views, result primitives based on the plurality of identified primitives from multiple views based on the plurality of identified entities.

A method and an apparatus for detecting a pedestrian by a vehicle during night driving are provided, in which the apparatus includes: a first camera configured to take a first image including color information of a vicinity of the vehicle during night driving; a second camera configured to take a second image including thermal distribution information of the vicinity of the vehicle; a pedestrian detector configured to detect a non-pedestrian area by using the color information from the first image and detect a pedestrian area by excluding the non-pedestrian area from the second image; and a display configured to match and display the pedestrian area on the second image.

A viewing device for a vehicle, comprises: a plurality of imaging apparatuses that differ in imaging range; an image generation apparatus that, on the basis of images imaged by one or a plurality of the imaging apparatuses, generates a viewing image that is a composite of: an image to a vehicle rear, an image of a vehicle structure, the image being semi-transparent, and an image of a vehicle occupant on a rear seat; and a displayer that displays the viewing image.

Systems and methods for automatically self-correcting or correcting in real-time one or more neural networks after detecting a triggering event, or breaching boundary conditions are provided. Such a triggering event may indicate incorrect output signal or data being generated by the one or more neural networks. In particular, machine controllers of the invention limit the operations of neural networks to be within boundary conditions. Autonomous machines of the invention can be self-corrected after a breach of a boundary condition is detected. Autonomous land vehicles of the invention are capable of determining the timing of automatic transition to the manual control from automated driving mode. The controller of the invention filters and saves input-output data sets that fall within boundary conditions for later training of neural networks. The controllers of the invention include security architectures to prevent damages from virus attacks or system malfunctions.

An image processing apparatus for processing an image according to one embodiment includes a generating unit, a display control unit, and a change receiving unit. The display control unit generates a synthetic image providing a view of a vehicle from a virtual viewpoint, based on a plurality of onboard camera images. The display control unit displays the generated synthetic image on a display unit. The change receiving unit receives a change in the relative positional relation between an image region that is based on one of the camera images and image regions that are based on the other camera images in the synthetic image. The generating unit generates a synthetic image, based on a changed positional relation every time a change in the positional relation is received.

A lighting apparatus that irradiates light on a vehicle is provided. The lighting apparatus includes: a first light source that emits light for illuminating a predetermined area; a memory that stores location information indicating a location of the lighting apparatus; a second light source that outputs, to the predetermined area, a light signal indicating the location information; and a first controller that controls the first light source and the second light source, and modulates the light signal output by the second light source, in accordance with the location information stored in the memory.

Example implementations may relate to use of a light-control feature to control extent of light encountered by an image capture device of a self-driving vehicle. In particular, a computing system of the vehicle may make a determination that quality of image data generated by an image capture device is or is expected to be lower than a threshold quality due to external light encountered or expected to be encountered by the image capture device. In response to the determination, the computing system may make an adjustment to the light-control feature to control the extent of external light encountered or expected to be encountered by the image capture device. This adjustment may ultimately help improve quality of image data generated by the image capture device. As such, the computing system may operate the vehicle based at least on image data generated by the image capture device.

An object detection device includes: an imaging unit configured to image the surroundings of a vehicle; a horizontal edge extraction unit configured to extract horizontal edges that are characteristic lines of approximately horizontal direction components in a search region of an image captured by the imaging unit; and a detection object recognition unit configured to recognize a detection object within a recognition region set on the basis of, among the horizontal edges extracted by the horizontal edge extraction unit, a specific horizontal edge satisfying a predetermined condition.

An apparatus includes a camera, a sensor and a processor. The camera may generate a video signal based on a targeted view of a driver. The sensor may generate a proximity signal in response to detecting an object within a predetermined radius. The processor may determine a location of the object with respect to the vehicle, determine a current location of eyes of the driver, determine a field of view of the driver at a time when the proximity signal is received based on the current location of the eyes, determine whether the object is within the field of view using the current location of the eyes, and generate a control signal. The distance may be determined based on a comparison of reference pixels of a vehicle component in a reference video frame to current pixels of the vehicle component in the video frames.

Systems and method for a driver assistance system including a surround view system are provided. In an example method for automatically selecting a virtual camera position in the surround view system, the method includes selecting one of the one or more vehicle surrounding the host vehicle as a threat vehicle based on at least one of a geographic position, and a velocity of the vehicle relative to one or more of a position, a heading, and a speed of the host vehicle. Based on the selected threat vehicle, the method includes selecting a virtual camera position such that the threat vehicle and a portion of the host vehicle are in view of a virtual camera, and displaying an image from the virtual camera position to a driver of the host vehicle.