|
|
|
| Moving object detection method based on complex fuzzy logic system |
| 1.School of Information and Electrical Engineering, Xuzhou Institute of Technology, Xuzhou, Jiangsu 221111, China; 2.School of Civil Engineering and Transportation, South China University of Technology, Guangzhou, Guangdong 510640, China |
|
|
|
|
Abstract To improve the accuracy, antijamming capability and efficiency of background modeling in complex scenes, the moving object detection method was proposed based on complex fuzzy logic system. The complex TakagiSugenoKang (TSK) fuzzy logic system was used as the estimation of background model, and a hybrid learning method combining particle swarm optimization (PSO) and kernel least mean square (KLMS) was used to learn the complex fuzzy logic system. The foreground pixels were regarded as outliers relative to the background ones, and an outlier separator method was proposed to train the complex fuzzy logic system. The foreground pixels were distinguished according to the comparison between estimated background model and foreground image. To verify the proposed method, the video sequences under three scenes of campus, highway and water surface were determined, and the detection performance was compared with those of other three classical methods. The results show that by the proposed method, high accuracy can be obtained under the conditions of dynamic background，illumination changes and camera vibration. The similarity measure value is 0.1 higher than those of other three methods with detection speed of 22 f·s-1, which can meet real time requirements.
|
|
Received: 30 August 2016
|
|
|
|
|
|
|
2017, Vol. 38 
