基于机器学习反馈的黑暗环境车辆路况自动化识别方法

Automatic Recognition of Vehicle Road Conditions in Dark Environment Based on Machine Learning Feedback

传统的车辆路况识别方法存在识别准确率低的问题,威胁驾驶员的人身安全。为此提出基于机器学习反馈的车辆路况识别方法。提取图像特征时利用函数分解窗将设定的像素函数分解出像素值,运用小波变换得出像素特征。使用极坐标方程将像素特征预处理成为像素簇。将像素簇代入到示警单元中进行扫描,得出阴影环境的像素峰值。示警单元自动将像素峰值转化为信号,传递给驾驶员。故此完成黑暗环境的车辆路况的自动化识别。对传统方法与本文提出的方法进行实验,实验结果表明,传统方式的识别准确率为78.4%,而自动化识别方法平均识别准确率达到了98.1%,具有更高的识别准确率。

The traditional vehicle condition recognition method has the problem of low recognition accuracy,which threatens the driver’s personal safety.Therefore,a vehicle condition recognition method based on machine learning feedback is proposed.When extracting image features,the function decomposition window is used to decompose the set pixel function into pixel values,and the wavelet transform is used to get the pixel features.The polar coordinate equation is used to pre-process the pixel features into a cluster of pixels.The pixel cluster is substituted into the warning unit for scanning,and the peak value of the pixel in the shadow environment is obtained.The warning unit automatically converts the peak value of the pixel into a signal and transmits it to the driver.Therefore,the automatic recognition of vehicle road condition in dark environment is completed.The experimental results show that the recognition accuracy of the traditional method is 78.4%,while the average recognition accuracy of the automatic method is 98.1%,which has higher recognition accuracy.