The invention discloses an extraction method for a key frame in a road vehicle monitoring video. The extraction method comprises the following steps: firstly, extracting a sample image set of a collected road vehicle monitoring video; extracting a minimum enclosing rectangle image of a vehicle as a sample image source through foreground moving object detection; aiming at enclosing rectangle sequential images of each vehicle object entering and leaving a monitoring area, judging to enable the enclosing rectangle sequential images meeting the vehicle monitoring analyzing demand as positive sample images and judging to enable the enclosing rectangle sequential images which do not meet the demand as negative sample images, thereby acquiring a training sample image set used for training an AdaBoost classifier; extracting area features and integral channel features of the foreground vehicle object, training and acquiring the AdaBoost classifier; utilizing the trained classifier to classify the enclosing rectangle image sequence of the object vehicle entering and leaving the monitoring area; rating each frame image by the AdaBoost classifier; selecting the image with the highest score as the key frame. According to the extraction method, the effectiveness and accuracy of the key frame extraction of the road vehicle monitoring video are increased.