Automatic logo recognition is one of the key elements of vehicle information collection in an intelligent transportation system（ITS）.Considering the abundant marginal information about the vehicle logo

this paper applies the characteristic values of the Histograms of Oriented Gradients（HOG） and implements the fast vehicle logo detection and recognition with the help of Support Vector Machine（SVM）.Furthermore

it proposes an improved HOG

which achieves evident results in terms of the accuracy of vehicle logo detection and recognition.Large amounts of experimental data and applications in ITS show that this method is fairly robust and practically useful.