基于智能视觉的警戒区域人员跨界检测仿真

Simulation of Personnel Crossover Detection in Alert Area Based on Intelligent Visual

在警戒区域人员跨界检测过程中,人员在警戒区的附近时,相关人员具备一定的隐蔽性和遮挡性,在相关人员故意遮挡的情况下,判断越过边界的检测阀值的过程会迅速模糊化,出现识别确认跳变的情况,检测的鲁棒性迅速下降,引起虚警和误判的现象。提出一种优化的警戒区域人员跨界图像检测方法。将采集的原始跨界图像进行跨界特征提取,将提取的特征图像进行相关匹配运算,参照因式分解的相关理论,建立警戒区域人员跨界视觉检测模型,融合非线性滤波原理,设置视觉检测图像的像素,构建非线性滤波窗口序列,对该视觉检测图像模型中的主动干扰进行滤波处理,去除主动遮挡干扰,优化识别过程。实验结果表明,提出的采用智能视觉算法的警戒区域人员跨界检测方法精度提高明显。

An optimized image detection method of personnel crossover in alert area is put forward. The original crossover feature and the characteristics of the image correlation matching arithmetic are extracted, a visual inspection model of personnel crossover in alert area is set up referring to the related theory of factoring and nonlinear filtering fusion principle, the visual inspection of the image pixels is set up, the nonlinear filtering window sequence is built, the active interference on the visual inspection image model is filtered and removed, and the identification process is optimized. The experimental results show that the proposed algorithm can increase the detection accuracy obviously.