Abstract: A dynamic image stitching system for stitching images captured by cameras of a vision system for a vehicle includes a plurality of cameras disposed at the vehicle and having respective fields of view exterior the vehicle. The field of view of each side camera partially overlaps the field of view of a front camera and partially overlaps the field of view of a rear camera. Image data captured by the plurality of cameras is processed at a processor to detect an object present in an overlapping portion of the fields of view of two cameras. The processor stitches, via a stitching algorithm, image data captured by the cameras. The processor adjusts the stitching algorithm responsive to a determination of a difference between a characteristic of the feature as captured by one of the two cameras and the characteristic of the feature as captured by another of the two cameras.

Abstract: A method for dynamically calibrating without manual intervention a forward viewing vehicular camera with respect to its three rotational degrees of freedom includes disposing a camera at a vehicle and operating the camera to acquire multiple frames of image data as the vehicle is moving and steered through a plurality of steering angles. Feature points are determined in an image frame when the vehicle is steered within a respective range of steering angles and motion trajectories of those feature points are tracked in subsequent image frames for the respective range of steering angles. A vanishing point is established in the image plane for the feature points for the respective range of steering angles. Based on established vanishing points, a vanishing line is determined. When the vanishing line is determined to be non-horizontal, at least one of pitch, roll or yaw of the camera is adjusted.

Abstract: A dynamic image stitching system for stitching images captured by cameras of a vision system for a vehicle includes a plurality of cameras disposed at the vehicle and having respective fields of view exterior the vehicle. The field of view of each side camera partially overlaps the field of view of a front camera and partially overlaps the field of view of a rear camera. Image data captured by the plurality of cameras is processed at a processor to detect an object present in an overlapping portion of the fields of view of two cameras. The processor stitches, via a stitching algorithm, image data captured by the cameras. The processor adjusts the stitching algorithm responsive to a determination of a difference between a characteristic of the feature as captured by one of the two cameras and the characteristic of the feature as captured by another of the two cameras.

Abstract: A method for dynamically correcting misalignment of a vehicular camera includes fixedly disposing a camera at a vehicle and operating the camera to acquire multiple frames of image data while the vehicle is moving generally in a straight line. A plurality of sets of feature points are selected in an image frame, with each set including a first feature point and a second feature point. For each set of feature points, a motion trajectory of that set's feature points is tracked in subsequent image frames. For each tracked first and second feature points, a vanishing point is established in the image plane. Based on the established vanishing point, a vanishing line is determined in the image plane. When the vanishing line is determined to be non-horizontal in the image plane, at least one of pitch, roll or yaw of the camera is adjusted to correct rotational misalignment of the camera.

Abstract: A dynamic image stitching system for stitching images captured by multiple cameras of a vision system of a vehicle includes a first camera disposed at a vehicle and having a first field of view exterior the vehicle and a second camera disposed at the vehicle and having a second field of view exterior the vehicle. The first and second fields of view at least partially overlap. A processor is operable to process image data captured by the first and second cameras. The processor processes captured image data to determine characteristics of features or objects present in the overlapping region of the first and second fields of view. The processor is operable to adjust a stitching algorithm responsive to a determination of a difference between a characteristic of a feature as captured by the first camera and the characteristic of the feature as captured by the second camera.

Abstract: A method for dynamically ascertaining alignment of a vehicular camera relative to a vehicle to which the camera is attached includes determining a plurality of steering angle ranges for the vehicle, each of which is a range of steering angles that approximates straight vehicle motion over less than two seconds of travel time of the vehicle while the vehicle is in motion and turning. Image data captured by a camera of the vehicle is processed to determine a central vanishing point when the vehicle is in motion and moving straight. A plurality of feature points are selected in the image frames for each steering angle range, and a vanishing point for a plurality of tracked motion trajectories for each steering angle range is determined. An alignment of the camera is determined based at least in part on the determined central vanishing point and a determined vanishing line.

Abstract: A method of dynamically calibrating a given camera relative to a reference camera of a vehicle includes identifying an overlapping region in an image frame provided by the given camera and an image frame provided by the reference camera and selecting at least a portion of an object in the overlapped region of the reference image frame. Expected pixel positions of the selected object portion in the given image frame is determined based on the location of the selected object portion in the reference image frame, and pixel positions of the selected object portion are located as detected in the given image frame. An alignment of the given camera is determined based on a comparison of the pixel positions of the selected object portion in the given image frame to the expected pixel positions of the selected object portion in the given image frame.

Abstract: A method for dynamically correcting misalignment of a vehicular camera includes fixedly disposing a camera at a vehicle and operating the camera to acquire multiple frames of image data whilst the vehicle is moving generally in a straight line. A plurality of sets of feature points are selected in an image frame, with each set including a first feature point and a second feature point. For each set of feature points, a motion trajectory of that set's feature points is tracked in subsequent image frames. For each tracked first and second feature points, a vanishing point is established in the image plane. Based on the established vanishing point, a vanishing line is determined in the image plane. When the vanishing line is determined to be non-horizontal in the image plane, at least one of pitch, roll or yaw of the camera is adjusted to correct rotational misalignment of the camera.

Abstract: A method of dynamically ascertaining the alignment of a vehicular camera. The method involves acquiring a sequence of images provided by the camera whilst the vehicle is in motion. For each range of steering angles, the method (i) selects a plurality of feature points in the images, (ii) tracks a motion trajectory for each selected feature point, and (iii) determines a vanishing point in the image plane based on the tracked motion trajectories. The method determines a vanishing line in the image plane based on a locus of these vanishing points and determines the alignment of the camera based on the position of a central vanishing point (corresponding to the zero degree angle) and the vanishing line.

Abstract: A method for calibrating a vehicular vision system includes providing a camera at a vehicle, with the camera having a field of view. Images are captured with the camera and a set of resultant images are acquired for a classification. Information is extracted related to image features in the set of resultant images, and an appropriate subset of coefficients is determined. For each classification, a classification vector of at least one appropriate weight is stored that corresponds to the determined subset of coefficients. The determined subset of coefficients is determined by processing sets of coefficients produced from a selection of calibration images and determining a subset of coefficients which acceptably discriminate between defined classifications. A set of resultant images is acquired by limiting the dynamic range of acquired images to obtain resultant images that include at least one region of interest.

Abstract: A method for calibrating a vehicular vision system includes providing a camera at a vehicle, with the camera having a field of view. Images are captured with the camera and a set of resultant images are acquired for a classification. Information is extracted related to image features in the set of resultant images, and an appropriate subset of coefficients is determined. For each classification, a classification vector of at least one appropriate weight is stored that corresponds to the determined subset of coefficients.

Abstract: A method of dynamically calibrating a given camera relative to a reference camera of a vehicle includes identifying an overlapping region in an image frame provided by the given camera and an image frame provided by the reference camera and selecting at least a portion of an object in the overlapped region of the reference image frame. Expected pixel positions of the selected object portion in the given image frame is determined based on the location of the selected object portion in the reference image frame, and pixel positions of the selected object portion are located as detected in the given image frame. An alignment of the given camera is determined based on a comparison of the pixel positions of the selected object portion in the given image frame to the expected pixel positions of the selected object portion in the given image frame.

Abstract: A dynamic image stitching system for stitching images captured by multiple cameras of a vision system of a vehicle includes a first camera disposed at a vehicle and having a first field of view exterior the vehicle and a second camera disposed at the vehicle and having a second field of view exterior the vehicle. The first and second fields of view at least partially overlap. A processor is operable to process image data captured by the first and second cameras. The processor processes captured image data to determine characteristics of features or objects present in the overlapping region of the first and second fields of view. The processor is operable to adjust a stitching algorithm responsive to a determination of a difference between a characteristic of a feature as captured by the first camera and the characteristic of the feature as captured by the second camera.

Abstract: A method for calibrating a vehicular vision system includes providing a camera at a vehicle, with the camera having a field of view. Images are captured with the camera and a set of resultant images are acquired for a classification. Information is extracted related to image features in the set of resultant images, and an appropriate subset of coefficients is determined. For each classification, a classification vector of at least one appropriate weight is stored that corresponds to the determined subset of coefficients.

Abstract: A method of dynamically ascertaining the alignment of a vehicular camera. The method involves acquiring a sequence of images provided by the camera whilst the vehicle is in motion. For each range of steering angles, the method (i) selects a plurality of feature points in the images, (ii) tracks a motion trajectory for each selected feature point, and (iii) determines a vanishing point in the image plane based on the tracked motion trajectories. The method determines a vanishing line in the image plane based on a locus of these vanishing points and determines the alignment of the camera based on the position of a central vanishing point (corresponding to the zero degree angle) and the vanishing line.

Abstract: A vision system for a vehicle includes a plurality of cameras disposed at a vehicle equipped with the vision system. Each of the cameras has a respective field of view and is operable to capture respective image data. The vision system may include or utilize a video file format that makes available required or desired image data and other information in a synchronized way, and that enables access to the image data and information by algorithms. The video file format may create a layout that allows the system to store and access the data in a generalized manner. The system may include a video control interface software tool that defines a protocol for interfacing with external devices and subsequent processing software.

Abstract: A vehicle interior classification system and method determines a classification relating to the interior of the vehicle, such as the occupancy status of a vehicle seat or the state of alertness of a vehicle driver, from one or more images of an appropriate portion of the interior of the vehicle acquired with an image capture device. The acquired images may be processed to limit the dynamic range of the images to obtain a resultant image that includes one or more regions of interest which are less than the total field of view of the image capture device. The resultant images are processed to extract information about features in the image. The set of coefficients produced with suchprocessing can be reduced to a subset of the total number of coefficients, the members of the subset being selected for their ability to discriminate between the classifications defined for the system.

Abstract: A vehicle interior classification system and method in accordance with the present invention determines a classification relating to the interior of the vehicle, such as the occupancy status of a vehicle seat or the state of alertness of a vehicle driver, from one or more images of an appropriate portion of the interior of the vehicle acquired with an image capture device. The acquired images are preferably processed to limit the dynamic range of the images to obtain a resultant image which can comprise one or more regions of interest which are less than the total field of view of the image capture device. The resultant images are processed to extract information about features in the image and, in one embodiment, this processing is achieved with a two-dimensional complex discrete wavelet transform which produces a set of coefficients corresponding to the presence and/or location of the features in the resultant image.

Abstract: A system and method for processing acquired images to develop useful classifications of subjects such as occupants of a vehicle preferably employs a hierarchical and probabilistic structure, such as a Bayesian Network to analyze acquired images and produce a meaningful classification. The structure preferably includes set of analyzers, a set of Scenario analyzers and a set of Temporal models which are arranged in three respective hierarchical layers. Each respective analyzer operates on the acquired image and, in some circumstances, feedback from the Scenario analyzers, to produce an output representing the probability that a feature that the respective analyzer is concerned with is present in the acquired image.