| 摘要 |
A method and system for training a vehicle detection system used in a street occupancy estimation of stationary vehicles. First and second areas are defined on an image plane of an image capture device associated with monitoring for detection of vehicles. Candidate frames of video-data that include objects relevant to a classification task in the second region, 44, are identified and objects are extracted from said images. Features of each object are extracted and labels are assigned to each extracted object. At least one classifier is then trained using the labels and extracted features. The at least one trained classifier is then utilised to classify a stationary vehicle detected in the first region, 42. The assigned label may be based on an attribute. The attribute may be one of size, shape, orientation or aspect ratio. The first area may be an on-street parking lane (parking bay, parking space) and the second area may be a lane of traffic adjacent to the parking area. |
