A method for estimating the speed of a vehicle moving in a single traffic lane is proposed. The blurring of the vehicle image and the presence of Gaussian noise at individual stages of image acquisition were taken into account. The method is based on the cross-correlation function determined between the distorted image and the vehicle model. The performance of the algorithm under conditions of strong random interference was verified through computer simulations.In this paper, the issue of vehicle speed estimation based on an image recorded by a video camera is considered. In the case of a low level of interference, a method consisting of subtracting consecutive image frames is used, which allows for the identification of moving objects. An example is the algorithm proposed by Lucas B.D and
Kanade T. [2]. One version of this algorithm involves the use of the image derivative.
Performing the signal differentiation operation in the presence of random interference can cause large errors [3]. In the case of a high level of interference, the method of subtracting consecutive image sequences becomes useless because it leads to an increase in the noise level – the noise variance is summed. (…)
The article contains 15075 characters.
Source: Logistyka Magazine