In order to overcome the problems described above,
this paper proposes an improved algorithm of video
compression for lane surveillance system. This new
algorithm is based on detection of moving vehicle and
vehicle-license-plate locating. First of all, we obtain
shot segmentation from the uncompressed video,
which is much fit for the character of the lane
surveillance. Second, the key frame, which mostly
represents the information of the shot segment, is
extracted. Finally, we use the result of shot
segmentation and key frame extraction to compress the
video efficiently. In detail, we directly use the key
frame as I frame, the former one as B frame, and the
latter one as P frame. Certainly, the efficiency of this
kind of video compressing method is higher than that
of other compressing method.
In order to overcome the problems described above,this paper proposes an improved algorithm of videocompression for lane surveillance system. This newalgorithm is based on detection of moving vehicle andvehicle-license-plate locating. First of all, we obtainshot segmentation from the uncompressed video,which is much fit for the character of the lanesurveillance. Second, the key frame, which mostlyrepresents the information of the shot segment, isextracted. Finally, we use the result of shotsegmentation and key frame extraction to compress thevideo efficiently. In detail, we directly use the keyframe as I frame, the former one as B frame, and thelatter one as P frame. Certainly, the efficiency of thiskind of video compressing method is higher than thatof other compressing method.
การแปล กรุณารอสักครู่..
