Fig. 1. The algorithm flow chart

Fig. 2. The mode background modeling process：（a）input image sequence，（b）grayscale change curve，（c）gray histogram

Fig. 3. Background modeling results of different algorithms

Fig. 4. Original MBLBP of frame 265 and frame 266

Fig. 5. The schematic diagram of S_MBLBP：（a）improved S_MBLBP，（b）S_MBLBP with a target

Fig. 6. The foreground target extraction：（a）gray value change diagram and gray histogram，（b）LBP value change diagram and LBP histogram，（c）foreground extraction process，（d）LBP background modeling

Fig. 7. Comparison results between the proposed algorithm and various LBP algorithms：（a）comparison of different LBP algorithms in Skating，（b）comparison of moving object extraction results from different LBP algorithms in Skating，（c）comparison of different LBP algorithms in Lakeside，（d）comparison of moving object extraction results from different LBP algorithms in Lakeside

Fig. 8. Processing results of different algorithms in complex scenarios：（a）skating，（b）snowfall，（c）overpass，（d）canoe

Fig. 9. The processing results of different algorithms in the infrared scene：（a）dining room，（b）lake side，（c）library，（d）park，（e）corridor

Fig. 10. Comparison of different algorithms in 9 scenarios

Table 1. Average processing time of visible light datasets

Table 2. Average processing time of infrared datasets

Table 3. Evaluation index results of different algorithms