基于图像识别轮胎变形的非接触式车辆称重方法

为了解决接触式车辆称重方法存在的安装和维修成本高、使用年限短、识别精度低等问题,创新性地提出一种基于计算机视觉获取轮胎变形的非接触式车重识别方法。首先,利用视频图像采集装置拍摄车辆轮胎图像信息,通过图像处理技术提取轮胎轮廓,并根据轮廓变形计算轮胎的垂向挠度。其次,通过胎压监测系统(TPMS)获取轮胎的真实胎压值,对于没有安装TPMS的车辆,则可以通过图像字符识别技术读取轮胎侧壁的胎压标识信息,再利用统计回归方式确定实际胎压值。在此基础上,将轮胎垂向挠度和胎压值代入推导的称重公式计算轮胎承受的重量,再将所有轮胎承受重量求和得到车辆总重量。最后,以现场的乘用车和重载货车为例,验证在不同胎压和重量变化下非接触式车辆称重方法的准确性,并对比分析3个称重公式的准确性。研究结果表明:车重识别准确率随着胎压增大而降低,随着车重增大而上升;轮胎刚度拟合公式的载重识别准确率达到95%以上,高于理论推导公式和半经验拟合公式。提出的非接触式车辆称重方法具有测量范围广、无需任何额外传感设备、不用封闭交通和易于信息集成等优势,有效地突破了现有接触式车重识别技术的瓶颈,具有很好的工程应用前景。

Non-contact Vehicle Weighing Method Based on Tire Deformation Using Image Recognition

Traditional contact-based weighing methods for moving vehicles have the disadvantages of high installation and maintenance costs,short service life,and low identification accuracy.Thus,a non-contact vehicle weighing method based on tire deformation using computer vision is proposed.First,a video/image acquisition device is used to obtain the tire image,and image processing techniques are applied to extract the tire edge and calculate the vertical deflection of the tire accordingly.Second,the tire pressure is obtained from the tire pressure monitoring system (TPMS).For tires without a TPMS,the actual tire pressure is obtained based on the pressure label information printed on the tire sidewall using image character recognition.Then,the obtained values of tire vertical deflection and tire pressure are inserted into a weighing formula to calculate the weight sustained by each tire,and the total weight of the vehicle is then acquired by summation.A passenger vehicle and a heavy-duty truck are used in a field test to verify the accuracy of the proposed method under different tire pressures and vehicle weights,and to compare the effectiveness of three different weighing formulas.The results show that the vehicle weight identification accuracy decreases with the increasing tire pressure,and increases with the increasing vehicle weight.The tire stiffness fitting formula is more accurate than the theoretically derived formula and the semi-empirical fitting formula,and the overall accuracy of vehicle weight identification exceeds 95%.The proposed non-contact vehicle weighing method has the advantages of a wide measurement range,no need for additional sensing equipment,no need for road closures,and easy integration,effectively overcoming the disadvantages of existing contact-based vehicle weighing technology and representing an excellent prospect for various engineering applications.