面向电力智能安监的极低分辨率目标检测算法

Very low-resolution object detection algorithms for electric intelligent safety supervision

为检测电力作业人员是否穿戴安全装备,提出一种面向电力智能安监的极低分辨率目标识别算法。在面对安全帽、护目镜等极低分辨率的目标检测时,当前目标检测算法无法保证检测精度。通过稀疏表示模型对极低分辨率下的目标图像进行超分辨率增强,对增强图像进行检测,对是否佩戴安全帽、护目镜等行为进行判别,满足电力智能安监的新需求。实验结果表明,该方法比当前最佳目标检测算法具有更高的检测精度,可以达到智能监控系统的实时性要求。

To solve the potential security problems of power system security caused by the workers unwearing the safety equipment,a very low-resolution object detection algorithm for power intelligent security monitoring was proposed.An algorithm based on sparse representation was proposed to detect small objects such as safety helmet and goggles.The very low-resolution objects were upscaled by sparse representation,the upscaled images were fed into the detectors and the abnormal behaviors were discriminated.Experimental results indicate that the proposed method outperforms the state-of-the-art methods,and can meet the real-time requirements of intelligent monitoring system.