Abstract: In order to verify the application effect of the Kalman filter algorithm in the detection of the smart car track, the Infineon TC264 microcontroller is used as the main controller, and the grayscale image of the track is collected through the camera. According to the pixel matrix of the black and white image, combined with software programming and proportion integration differentiation (PID) algorithm, the duty cycle of the motor and the steering gear is calculated, and the forward direction of the car is correctly controlled, thereby realizing the tracking function. After the car can track normally, the Kalman filter algorithm is embedded in the original algorithm, and the specific application effects of the two algorithms on the car are compared. After continuous debugging, the car can realize the tracking function, the speed can reach 1.4 m/s, and after adding the Kalman algorithm, the car can predict the track ahead. The experimental results show that the Kalman filter algorithm can optimize the track recognition.

Key words: Camera tracking; Image binarization; Pulse width modulation; Kalman filter

摘要： 为了验证卡尔曼滤波算法在摄像头智能车赛道检测方面的应用效果，使用英飞凌 TC264 单片机为主控制器，通过摄像头采集赛道的灰度图像。根据图像的像素矩阵结合软件编程和 比例-积分-微分（PID）算法，计算出电机和舵机的占空比，正确控制摄像头智能车的前进方 向，从而实现循迹功能。在摄像头智能车能够正常循迹之后，将卡尔曼滤波算法嵌入到原有 算法中，比较前后两种算法在摄像头智能车上的具体应用效果。经过调试，摄像头智能车可 以实现循迹功能，速度可达 1.4 m/s，且加入卡尔曼滤波算法之后，摄像头智能车能够预测前 方赛道。试验结果表明，卡尔曼滤波算法可以对赛道识别起到优化作用。

关键词: 摄像头循迹；图像二值化；脉冲宽度调制；卡尔曼滤波算法