Taxi detection method based on Haar-Adaboost algorithm and histogram comparison

The invention discloses a taxi detection method based on a Haar-Adaboost algorithm and histogram comparison. Firstly, positive and negative samples in a window area are acquired, and Haar features ofthe positive and negative samples are extracted; then, the positive and negative samples are processed through Adaboost algorithm-based cascade classifier training, a to-be-detected image and a videoand samples trained after the Adaboost algorithm-based cascade classifier are inputted to a loading cascade classifier, a taxi detection result is displayed, a roof light area is positioned in the detected area, and color space conversion is carried out in the roof light area; and finally, histogram comparison is carried out, a taxi mutual blocking condition which is most likely to happen in an actual environment is considered, and a window area of the taxi which is not likely to be blocked is selected as a Haar feature extraction area. In consideration of real-time detection which needs to berealized for a taxi in a c