局部纹理特征改进空闲车位智能检测算法

针对车辆数量的爆炸式增长而智能停车场车位检测速度慢、正确率低的问题,提出了一种基于局部二值纹理特征改进的空闲车位检测算法,算法首先将监控摄像机采集的车位图像降噪滤波后转化为局部二值模式,然后通过结构模板集提取图像的局部二值纹理特征,作为空闲车位和有车车位的识别特征,最后通过核Fisher判别方法实现空闲车位的检测。实验结果表明,所提方法对空闲车位的正确检测概率达到91.57%,对有车车位的错误检测概率只有4.46%,满足停车场空闲车位自动检测的要求,检测性能优于实验中采用的对比方法。

Vacant Parking Cell Detection Based on Spatial Texture Histogram Features

Aiming at the problem of high complexity and low accurate rate in vacant parking cell automatic detection of smart park,a detection method based on spatial texture histogram features is proposed.Firstly,we transform the pre-processed parking cell image to local binary pattern(LBP)image.And then,the spatial texture histogram feature,which has excellent performance of discriminating vacant parking cell and car parking cell,is extracted based on structure features.Finally,the kernel Fisher is used to detect vacant parking cell.The experimental results show that the proposed method is superior to other algorithms.Its correct detection probability of vacant parking cell is 91.57%,error detection probability of car parking cell is only 4.46%.The performance of proposed method satisfies the requirement of vacant parking cell detection in smart park.