Convolution algorithms present a key component and a significant step in image processing field. Despite their high arithmetic complexity, these algorithms are widely used because of their great importance for extracting image properties and features. Convolution algorithms require significant computing time, for that we propose a GPU acceleration of these algorithms by using the programming language CUDA presented by NVIDIA. Since these algorithms consume a lot of computing power, we understand the impact of the implementation of this type of algorithm on the acceleration of processing. GPU implementation present a suitable path to achieve better results than other  implementation , for that optimizing time consuming time consuming of applications became an increasingly important task in many research areas. The goal of this work is to try to boost convolution algorithms execution time by adopting GPU implementations to accelerate treatments and to achieve real time constraints.

In the last two decades, developing Driving Assistance Systems for security has been one of the most active research fields in order to minimize traffic accidents. Vehicle detection is a vital operation in most of these applications. In this paper, we present a high reliable and real-time lighting-invariant lane collision warning system. We implement a novel real-time vehicles detection using Histogram of Oriented Gradient and Support Vector Machine which could be used for collision prediction. Thus, in order to meet the conditions of real-time systems and to reduce the searching region, Otsu’s threshold method play a critical role to extract the Region of Interest using the gradient information firstly. Secondly, we use Histogram of Oriented Gradient (HOG) descriptor to get the features vector, and these features are classified using a Support Vector Machine (SVM) classifier to get training base. Finally, we use this base to detect the vehicles in the road. Two sets generated the training data of our system a set of negative images (non-vehicles) a set of positive images (vehicles), and the test is performed on video sequences on the road. The proposed methodology is tested in different conditions. Our experimental results and accuracy evaluation indicates the efficiency of your system proposed for vehicles detection.