Comparative analysis of several vehicle detection methods in urban traffic scenes

Improved license plate localisation algorithm based on morphological operations

“…Another texture detecting technology uses Haar-like methods for object detection [22], [23], [26]. Haar-like methods classify size, colour, brightness and location of license plates thereby aiding in the detection of characters on a plate.…”

“…However, a major drawback of Gabor filters is that they are time consuming. Another popular method is wavelet transform and is based on small wavelets with limited duration [2,6,16]. In this method, vertical features are extracted using wavelet transform and the position parameters of the plate are determined by analysing the projection features in both the time and frequency domains.…”

“…In this method, vertical features are extracted using wavelet transform and the position parameters of the plate are determined by analysing the projection features in both the time and frequency domains. Another texture detecting technology uses Haar-like methods for object detection [2,15,17]. Haar-like methods classify size, colour, brightness, and location of license plates thereby aiding in the detection of characters on a plate.…”

“…Automatic license plate recognition (ALPR) is a computer technology where an optical device captures an image and this is then processed to obtain a string from a license plate. In general, the ALPR systems require a high-performance workstation to work [1][2][3]. However, when the same algorithm is used on a low computational power device, the processing time is extremely high relative to a workstation and it takes many seconds to achieve a result.…”