Sports Target Tracking and Speed Measurement Method Based on Time Domain Vision Sensor

Moving target tracking technology is an important topic in the field of time domain sensors, which widely used medical diagnosis, intelligent robots, human-computer interaction, education, and entertainment. This article based on vision sensor to explore methods sports speed measurement. The introduces regional gray-scale correlation matching method algorithm, noise model sensor, eliminating noise. At same time, effects SIFT Mean Shift MSO algorithm moving targets are studied, analyzing ANF TSML IHTLS visual compare TUD-Crossing test set, ETHBAHNHOF SMOT data set; then image was purified analyzed sensor. research results show that, exploring when improved uses feature adaptive fusion describe target, it can automatically allocate proportion color features texture features, improves accuracy reliability tracking. Among them, DNP control final rate above 80%, its recall basically greater than 80%. In case that rough less 50%, after processing by multiple purification algorithms 98%.