基于可变形部件模型的安全头盔佩戴检测

利用视频监控系统进行安全头盔佩戴检测，对于安全生产有着重要意义。已有的安全头盔佩戴检测算法有较多的应用场景条件限制，难以同时满足不同场景需求。针对这一问题，提出了一种应用于安全头盔佩戴检测的算法。该方法依托可变形部件模型，提出了基于块的局部二值模式直方图，与梯度方向直方图和颜色特征共同组成特征向量，使用支持向量机进行训练和检测，利用了单一使用梯度方向直方图作为特征时所损失的有效信息。实验结果表明：该方法优于原可变形部件模型，在安全头盔测试集上的平均检测率提高了7.2%，达到86.7%，已接近应用要求。

A method for detecting safety helmet based on deformable part model

Safety helmet wearing detection through video monitoring system is extremely meaningful for safety production management. However, the present detection algorithms have too many application restrictions to meet the needs of various occasions. Therefore, this paper developed an improved algorithm to solve the problem. A modified deformable part model (DPM) was proposed by a combination of histogram of block-based local binary pattern (HOB-LBP) (which is based on local binary pattern), histogram of oriented gradient (HOG) and color features. With the usage of support vector machine for training and detecting, our algorithm effectively diminished the loss of information compared with the original DPM which only included the HOG feature. According to the experiment on the safety helmet database, the average precision has been increased by 7.2% and reached 86.7%, meeting the application requirements basically.