Novel pedestrian multi-target tracking method in complex traffic scenarios

算法1：自适应遗忘卡尔曼滤波算法（FSA）
输入：初始化测量矩阵 $z_{t}$ 、测量值的置信度 $c_{t}$ 、预测均值 ${\hat{x}}_{\bar{t - 1}}$ 、方差 $P_{\bar{t - 1}}$ 、观测矩阵 $H_{t}$ 、噪声协方差 $R_{t}$ 、卡尔曼增益 $K_{t}$ 和自适应因子 $μ$
输出：输出最终的预测均值 $t$ 和方差 $P$
1 预测目标状态；
2 跟踪与检测结果相匹配；
3 根据匹配和检测结果更新 $y_{t} = z_{t} - H_{t} {\hat{x}}_{\bar{t - 1}}^{'}$ ；
4 测量噪声协方差 ${\tilde{R}}_{t} = [(1 - c_{t}) \times \frac{1}{k} \sum_{t = 1}^{k} c_{t}] \times R_{t}$ 用于表示测量噪声尺度；
5 计算卡尔曼增益 $K_{t}$ ；
6 更新参数， $\{\begin{array}{l} {\hat{x}}_{\bar{t}} = {\hat{x}}_{\bar{t - 1}}^{'} + K_{t} y_{t} \\ P_{\bar{t}} = (\frac{1}{μ}) (I - K_{t} H_{t}) P_{\bar{t - 1}}^{'}, μ = 0.95 \end{array}$ 。

Table 1. Partially trained weights for YOLOv8

View table

View in Article

Table 1. Partially trained weights for YOLOv8

层序号	模块/层	输入通道数	输出通道数	卷积核大小	步长	填充
0	Conv	3	16	3×3	2	1
1	Conv	16	32	3×3	2	1
2.0	Conv（cv1）	32	32	1×1	1	0
2.1	Conv（cv2）	48	32	1×1	1	0
4.0	Conv（cv1）	256	256	1×1	1	0
4.1	Conv（cv2）	384	256	1×1	1	0
5.0	Conv（cv1）	256	128	1×1	1	0
5.1	Conv（cv2）	512	256	1×1	1	0

Table 2. Performance of YOFGD model on MOT16

View table

View in Article

Table 2. Performance of YOFGD model on MOT16

算法	MOTA↑	IDF1↑	MT↑	FPS↑
YOFGD	69.7	71.3	41.5	55.8
YOLOv8+CNN+DeepSORT	61.6	62.4	34.9	57.5
JDE-864	62.1	56.9	34.4	24.1
JDE-1088	64.4	55.8	35.4	18.9
FairMOT	68.7	70.4	39.5	25.9
Tracktor	54.5	52.5	19.0	5.0
CenterTrackPub*	67.6	57.2	32.9	6.8

Table 3. Performance of YOFGD model on MOT17

View table

View in Article

Table 3. Performance of YOFGD model on MOT17

算法	MOTA↑	IDF1↑	MT↑	FPS↑
YOFGD	77.9	70.6	45.7	48.7
YOLOv8+CNN+DeepSORT	67.5	55.9	36.4	56.3
JDE-864	66.6	57.2	32.9	19.4
JDE-1088	67.6	57.4	32.2	18.5
FairMOT	67.5	69.8	37.7	18.9
Tracktor	56.4	52.3	19.5	25.0
CenterTrackPub*	61.5	53.3	26.4	18.0

Tools

Get Citation

Copy Citation Text

Wenshun SHENG, Jiahui SHEN, Qi CHEN. Novel pedestrian multi-target tracking method in complex traffic scenarios[J]. Chinese Journal of Liquid Crystals and Displays, 2025, 40(5): 785

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category:

Received: Aug. 16, 2024

Accepted: --

Published Online: Jun. 18, 2025

The Author Email: Wenshun SHENG (sws@njpji.edu.cn)

DOI:10.37188/CJLCD.2024-0238

Topics