基于KL变换和支持向量机的空闲车位检测方法

Free Parking Space Detection Based on KL Transform and SVM

空闲车位自动检测是现代智能化停车场设计时面临的重要问题,针对此问题,提出了一种基于KL变换和支持向量机的空闲车位自动检测方法,算法首先对经过灰度化和滤波处理后的车位图像进行KL变换,将车位图像从图像空间映射至特征子空间,从而提取出用于检测的特征参量;在此基础上利用训练后的支持向量机完成空闲车位的检测。实验结果表明,该方法能够有效检测出车位的停车信息,对噪声不敏感,且能够对不同环境采集的车位图像具有很强的适应性和鲁棒性。

The free parking space detection is the kernel issue in designing intelligent parking lot. To solve this problem,a free parking space detection based on KL transform and support vector machine is proposed. Firstly,the KL transform is applied to the parking space image which is pretreated by graying and filtering,based on which,the parking space image was mapped to feature subspace and characteristic parameters used for parking space detection detection was extracted. And then,the free parking space detection was achieved by trained support vector machine using above parameters. Experimental results show that,the proposed method can effectively detect parking information of parking spaces,be insensitivity to noise and has strong adaptability and robustness to the parking space images collected in different environments.