基于计算机视觉的逃逸车辆识别系统设计与研究

Based on Computer Vision Escape Vehicle Identification System Design and Research

针对在进行智能化的肇事逃逸车辆图像定位中,车辆行驶较为快速,车身所处背景变化存在较大的随机性。背景的颜色灰度等特征的变化非线性和突变性很强。很难建立较为稳定的车辆定位模型,定位准确性不高。为此,提出一种基于计算机动态图像处理技术的肇事逃逸车辆定位方法,通过进行动态性较强的图像处理手段,克服背景车辆快速变化的弊端,准确的定位行进中的车辆。实验表明,这种计算机动态视觉方法能够大幅提高肇事逃逸车辆图片的关键帧定位结果的准确性。

Aiming at the intelligent textbook hit-and-run images in positioning vehicles,vehicle relatively fast,the body in the background there exists much randomness in change.Background color gray characteristics of the nonlinear and change by the strong.It is difficult to create a stable vehicle locating model,positioning accuracy is not high.For this,put forward a kind of based on computer dynamic image processing technology textbook hit-and-run vehicle positioning method,the dynamic strong image processing method,to overcome the shortcomings of the rapidly changing background vehicles,accurate positioning the traveling vehicles.Experiments show that the computer dynamic visual method can greatly improve the spin-off vehicle picture key frames the accuracy of the locating results.