Research on the monitoring technology for geological disasters by the intelligent video recognition technology based on the optical flow

Abstract

As one of the integrated monitoring technology, the video recognition technology has been applied in the geological disaster monitoring. The urgent problems are to deal with the influencing factors such as the complex environmental changes, occlusion, jitter, etc.. After solving these problems, the difficulty of identifying and tracking the disaster body is reduced, and the physical information recognition of the target body and the acquired video image processing are more accurate. As the result, the capture, recognition and analysis of the motion characteristics of the disaster body can be realized. According this technology, the video image stabilization is used to do de-jitter pre-processing, and then the sparse optical flow algorithm and the dense optical flow algorithm to identify and analyse the image, and the motion information of the disaster body can be identified end. As an empirical example, the video of Zhangjiawan rock avalanches was dealed with this technology. This example shows that this geological hazard monitoring technology can more accurately identify the motion information of the disasters, and it is suitable for the difficult early identification of rockslides in the complex environment of the karst mountainous areas in Southwest China.