Proc. SPIE. 7283, 4th International Symposium on Advanced Optical Manufacturing and Testing Technologies: Optical Test and Measurement Technology and Equipment
KEYWORDS: Detection and tracking algorithms, Digital filtering, Digital imaging, Image filtering, Reconstruction algorithms, Intelligence systems, Electronic filtering, Binary data, Filtering (signal processing), RGB color model
In the transportation monitor system, motive vehicle detection by adopting digital image is one of key technologies. To
detect motive vehicle accurately, we establish an adaptive background updating model firstly. Noise is suppressed by
using modality filter, and we obtain binary image by using maximum entropy to choose dynamic adaptive threshold.
Based on positive information of shadow and aspect feature of motive vehicle, we adopt HSV colour space and double
threshold to solve the problem of vehicle shadow. According to prediction result of Kalman filtering; we set up the area
of interest (AOI) of the vehicle model and adjust the size of AOI dynamically in order to track vehicle accurately. The
results of experiment show that, the arithmetic proposed in the paper can suppress shadow availably, detect motive
vehicle accurately and satisfy real-time motive vehicle tracking.