基于模糊增强的安全带佩戴识别方法

摘要

To overcome the insufficency of existing warning systems for seat⁃belt wearing, and to further increase the seat⁃belt wearing rate, a new method for recognizing whether a seat⁃belt is worn or not is proposed based on fuzzy enhancement. In this method, firstly, the collected images of the driver are processed by light compensation. After using the customized Canny operator for edge detection, the straight⁃line detecting based on Hough transform and the preliminary judgment are proceeded. By processing the images in which the seat belt hasn’ t been detected, a picture with fuzzy enhancement is obtained, then the Otsu method is used to select the optimal threshold for image segmentation, deriving the binary images. And then the straight⁃line detection is proceeded and thus the final con⁃clusion is reached. The test results show that the method proposed can effectively recognize the seat⁃belt wearing state with high accuracy.%为了弥补目前安全带佩戴提示系统的不足，提高驾驶员安全带佩戴率，提出一种基于模糊增强的安全带识别方法。该方法首先对采集到的驾驶员图像进行光补偿处理，采用自定义的Canny算子进行边缘检测，再使用Hough变换对处理过的图像进行直线检测和初步判断，对于没有检测到安全带的图像进行模糊增强，采用最大类间方差法选择最佳阈值进行图像分割得到二值化图，再进行直线检测并得到最后结论。实验证明，该方法有效地完成对驾驶员安全带佩戴情况的识别，具有较高的准确率。