基于云模型移动对象行为识别算法评价

Moving object behavior recognition algorithm evaluation based on cloud model

摘要:

为了研究视频监控中异常行为识别算法鲁棒性、准确度和速度的评价方法，将采集图像序列移动对象运动类型等参数作为样本输入逆向云发生器，得到移动对象定性概念的定量表示：期望值Ex、熵En和超熵He。以这些参数作为基础模拟出移动对象行为参数表征。Ex,En和He输入正向云发生器，将每个移动对象设计成一个智能体，个体通过感知环境和自激励调整行为参数，产生多种行为表征参数，并用这些参数来评价行为识别算法。通过实验实现了多种行为表征参数的模拟，用这些参数评价了几种典型算法，实验结果表明评价方法切实可行。

Abstract:

Abnormal behavior recognition algorithm evaluation of robustness, accuracy, and speed in video surveillance is proposed. Moving object parameters such as moving types in image sequence are collected and inputted into backward cloud generator. The quantitative representation of qualitative concepts, expectation Ex, entropy En, and super entropy are obtained. These parameters are adopted to simulate the moving object behavior representation parameters. Ex, En, and He are inputted into normal cloud generator. Each moving object is designed as agent that could adjust behavior parameters by sensing environment and auto excitation. These behavior parameters are adopted to evaluate some classics algorithm; the experiment results show that this evaluation methodology is effective and practical.