Real-time online traffic state detection method based on traffic monitoring video

The invention discloses a real-time online traffic state detection method based on a traffic monitoring video. The device is characterized in that the device comprises a video processor which is corresponding to each monitoring device, and each video processor is connected with a remote server through a network module. The method comprises the steps that each video processor reads a video image corresponding to a monitoring device; according to acquired pictures, the grayscale value X of each pixel in each picture is acquired, and is compared with a pre-stored parameter value X-1; a nuclear method is used to assess a road congestion extent ci observed from the video processor, and ci is uploaded; and a remote server analyzes each real-time road congestion extent ci to acquire a road congestion extent value C from one intersection to a next intersection. According to the invention, each video processor analyzes the real-time image grayscale; the nuclear method is used to assess and upload the congestion extent; the amount of data transfer is reduced; the server acquires data in real time; and the processing capacity and the effectiveness of traffic information are improved.