摘要: |
Vehicles may be recognized from the sound they emitwhen driving along a road. Characteristic acoustic fingerprints and audio features can be used to increase the robustnessof existing video based vehicle tracking and classificationalgorithms. Using this information in a multisensorsurveillance system helps to improve various parameterssuch as recognition rates, detection times and robustness.We propose a two-fold approach, where vehicle detectionand classification are handled separately. We demonstratethe feasibility of the proposed method using outdooraudio sequences of traffic situations. |