A raindrop recognition device is configured to recognize a raindrop on a transparent panel. The raindrop recognition device includes a storage unit, an image input unit, and an image recognition unit. The storage unit stores a trained model that is a machine learning model trained using, as training data, images of the transparent panel with adhered raindrops and images of the transparent panel without adhered raindrops. Image data of the transparent panel taken by a camera is input to the input unit. The image recognition unit is configured to calculate a value representing a raindrop likeness of an object on the transparent panel in the image data by inputting the image data to the trained model. The trained model is trained by training data including images of the transparent panel with uniform background and images of the transparent panel with light source.

An apparatus includes an interface and a processor. The interface may be configured to receive video frames comprising at least one of a vehicle or an outdoor environment near the vehicle and a correction signal. The processor may be configured to perform video operations on the video frames to detect objects, predict a state of the vehicle based on the objects detected in the video frames and generate the correction signal. The correction signal may be configured to apply a corrective measure based on the predicted state of the vehicle. The state of the vehicle may be predicted when the vehicle is parked in the outdoor environment. The state of the vehicle may comprise factors that prevent driving the vehicle.

A wiper is controlled to sweep a surface area, e.g., a portion of a windshield, in front of a camera, e.g., a camera mounted inside a vehicle. An image captured from the camera after the wiper completes, e.g., immediately completes, the sweep of the surface area in front of the camera is used in generating a depth map. In some embodiments a first wiper is controlled to clear a surface area in front of a first camera while a second wiper is controlled to clear a surface area in front of second camera at the same time, and the first and second cameras are synchronized to initiate image capture at the same time, capturing images through recently cleared area, and the captured images are used to generate a depth map. A vehicle control operation, e.g., a direction, braking or speed control operation is performed based on the depth map.

A windshield-wiper system includes a precipitation-detector, a windshield-wiper actuator, an object-detector, and a controller. The precipitation-detector detects precipitation proximate to a host-vehicle. The windshield-wiper actuator clears the precipitation from a windshield of the host-vehicle. The object-detector detects a distance of an object to the host-vehicle. The controller is in communication with the precipitation-detector, the windshield-wiper actuator, and the object-detector. The controller determines when the precipitation is present based on the precipitation-detector, determines the distance from the object to the host-vehicle based on the object-detector, and adjusts a speed of the windshield-wiper actuator when the precipitation is detected and the object is less than a distance-threshold away from the host-vehicle.

An occluded area on a vehicle sensor is identified by localizing the vehicle in space. A discrepancy between historic image data and real-time image data from the sensor is determined, and a cleaning actuator is actuated based on the determined discrepancy.

A system includes a processor and a memory. The memory stores instructions executable by the processor to detect an occlusion on a surface in a vehicle sensor optical path based on a segmentation of sensor image data, and to select a cleaning plan for the surface based on the detected occlusion, map data, and vehicle route data.

A system includes a sensor window, a sensor having a field of view through the sensor window, a piezoelectric vibrator positioned to impart vibrations to the sensor window, and a computer communicatively coupled to the sensor and the piezoelectric vibrator. The computer is programmed to identify a type for an obstruction of the sensor window based on data from the sensor, and instruct the piezoelectric vibrator to vibrate the sensor window with vibrations having a vibration profile. The vibration profile has at least one of frequency or phase velocity chosen according to the identified type of obstruction.

The invention relates to a cleaning system (14) for a windscreen (34) of a motor vehicle (32), comprising at least one drive arm (2) and at least one wiper blade (1, 1a, 1b), the wiper blade (1, 1a, 1b) being connected to the drive arm (2), the cleaning system (14) having at least one projection device (13) able to project a liquid onto the windscreen (34), characterized in that the cleaning system (14) comprises a determination element (16) for a zone (23) of the windscreen (34) and an activation means (15) of the determination element (16), an activation of the drive arm (2) and of the projection device (13) being subject to image information (18) as to a position of the zone (23) on the windscreen (34) transmitted by the determination element (16).

A windshield-wiper system includes a precipitation-detector, a windshield-wiper actuator, an object-detector, and a controller. The precipitation-detector detects precipitation proximate to a host-vehicle. The windshield-wiper actuator clears the precipitation from a windshield of the host-vehicle. The object-detector detects a distance of an object to the host-vehicle. The controller is in communication with the precipitation-detector, the windshield-wiper actuator, and the object-detector. The controller determines when the precipitation is present based on the precipitation-detector, determines the distance from the object to the host-vehicle based on the object-detector, and adjusts a speed of the windshield-wiper actuator when the precipitation is detected and the object is less than a distance-threshold away from the host-vehicle.

A system includes a processor and a memory storing instructions executable by the processor to predict a path of a liquid droplet on a surface, and then, actuate one or more vehicle components based on the path.