基于HSI模型和Hough变换的指针式汽车仪表自动校验

为了克服传统的汽车仪表人工校验普遍存在准确率不高的弊端，提出一种指针式汽车仪表自动校验算法.针对汽车仪表盘颜色信息丰富的特征和采集图像时目标物体周边有阴影的现象，将RGB模型下的表盘图像转到HSI（色调、饱和度、亮度）模型以消除阴影影响，并利用色调和饱和度分量作为特征参数进行图像分割，并采用Hough变换检测目标；分别计算出指针指向主要刻度线时相对于零刻度线的偏转角,并与相对应的标准偏转角比较，若超过允许误差范围则返回重调.对常用汽车仪表盘的自动校验结果表明该算法精度好，校验准确率高.

Automatic checking of pointer automotive dashboardbased on HSI model and Hough transformation

A novel approach for pointer automotive dashboard automatic checking was presented as the traditional manual checking method having a low accuracy. When taking images, the dashboard’s color information was high in abundance and there was a shadow around the object, so, the RGB model image was converted to HSI (Hue, Saturation, Intensity) model avoiding the influence of shadow. Then with the Hue and Saturation as two characteristic parameters, the image was segmented. After that, the objects were detected by means of the Hough transformation. The deflection angles, the movements of the pointer from the zero position, were calculated when the pointer pointing to the main marks. Comparing the deflection angle to the standard angle, the dashboard would be modified again if it was out of the error range. Experiments on the dashboard in the common use demonstrate that the proposed approach has a nice precision and increases the checking accuracy.