A neural network may be used to determine corner points of a skewed polygon (e.g., as displacement values to anchor box corner points) that accurately delineate a region in an image that defines a parking space. Further, the neural network may output confidence values predicting likelihoods that corner points of an anchor box correspond to an entrance to the parking spot. The confidence values may be used to select a subset of the corner points of the anchor box and/or skewed polygon in order to define the entrance to the parking spot. A minimum aggregate distance between corner points of a skewed polygon predicted using the CNN(s) and ground truth corner points of a parking spot may be used simplify a determination as to whether an anchor box should be used as a positive sample for training.

An image processing apparatus includes a conversion unit, a generation unit, and a processing unit. The conversion unit is configured to convert an original image received from a plurality of cameras into a top view image. The generation unit is configured to generate a surround view monitor (SVM) image by synthesizing a bottom image extracted from the top view image and a wall image extracted from the original image. The processing unit is configured to generate a modified SVM image by adjusting an area of the bottom image and an area of the wall image according to an adjustment condition, and to process the generated image into a display signal.

Provided is a method and system that includes an imaging device to be disposed in a lighting fixture to capture images, a remote computing device in communication with the imaging device, and comprising at least one processor. The processor is capable of processing data related to images from the at least one imaging device and from a satellite imagery system, performing comparison operation by detecting a first set of points of interest in an image from the at least one imaging device and corresponding second set of points of interest in an image of a same area from the satellite imagery system, and calculating a homography matrix by matching the first set of points of interests in the image from the at least one imaging device and the second set of points of interest in the image from the satellite imagery system, to determine latitude and longitude coordinates of the image from the imaging device.

Improvement in the accuracy of estimating the position of a mobile entity even while traveling or if there is an error in the calibration performed utilizing: a mobile entity; an imaging device provided in the mobile entity; and an information processing device for determining a first movement amount by which a detection point that is the same object has moved on the basis of a first image and a second image acquired by the imaging device and a second movement amount by which the mobile entity has moved while the first image and the second image were acquired, determining the accuracy of recognizing the detection point acquired by the imaging device on the basis of the first movement amount and the second movement amount, and estimating the position of the mobile entity on the basis of the accuracy of recognition and position information that pertains to the detection point.

An embodiment of the invention may include a method, computer program product and computer system for vehicle location detection. The method, computer program product and computer system may include a computing device which may receive image data from an imaging device associated with a vehicle and sensor data from a vehicle sensor device associated with the vehicle. The computing device may detect the vehicle has entered a parking scene based on the received image data. The computing device may detect the surroundings of the vehicle using the imaging device. The computing device may determine the vehicle is parking based on the received sensor data. The computing device may identify a parking location of the vehicle based on the received image data, receive image data and detect the surroundings associated with the identified parking location. The computing device may generate a notification to a user associated with the vehicle.

A camera parameter set calculation method including (a1) acquiring a first image captured by a first camera and a second image captured by a second camera, (a2) acquiring camera parameter sets of the first and second cameras, (a3) extracting, from the first image, a predetermined portion having assumed three-dimensional position information, (a4) calculating three-dimensional coordinate sets corresponding to the predetermined portion on a basis of the assumed three-dimensional position information, the first image, and the first camera parameter set, (a5) determining first pixel coordinate pairs and second pixel coordinate pairs corresponding to the predetermined portion in the first image and second image, respectively, on the basis of the three-dimensional coordinate sets, (a6) calculating an evaluation value on the basis of pixel values at the first pixel coordinate pairs and pixel values at the second coordinate pairs, (a7) updating the camera parameter sets on the basis of the evaluation value.

An image processing device is equipped with a marker detector, an oblique line detector, and a parking frame detector. The marker detector extracts, from an image, positive and negative edge segments that are aligned next to each other at a predetermined interval as a pair, and detects a parking space division marker based on the pair. The oblique line detector detects an edge on one side of an edge segment (single-edge segment) that is not extracted as the pair, and detects a presumed straight line based on the single-edge segment connected with an oblique line. The parking frame detector detects neighboring edge segments based on the positive and negative edge segments as well as the single-edge segment of the presumed straight line, and detects a parking space based on a distance between the neighboring edge segments.

Disclosed are methods and systems for providing pedestrian counts and providing the likelihood of parking at various locations on a street based on image data obtained from camera equipped vehicles imaging the street and the vehicle travels along the street.

Systems and methods to recognize a parking space are disclosed. One or more images are captured using one or more cameras coupled to a vehicle at a parking space. Data associated with the parking space are determined from the one or more images. The data associated with the parking space are compared to data stored in a storage device. A parking condition for the vehicle is determined based on the comparing.

A method to be implemented by a computer system includes: transforming each of training images into a respective transformed training image; calculating a classification loss value based on detection results that are acquired from parking status prediction results obtained by performing feature extraction on the transformed training images; and adjusting candidate spatial transforming parameters, candidate feature extraction parameters and candidate logistic regression parameters when it is determined that a combination of the aforementioned parameters is not optimal based on the classification loss value, followed by repeating above-mentioned steps using the adjusted parameters.