Landslide detection system using computer vision approach and Raspberry Pi

摘要

Landslide is a natural hazard, which badly affects people's safety and property. Continuous monitoring of such a disaster may lead to reducing of losses of human lives. With this aim, in this paper, we have proposed a surveillance system for real time landslide monitoring using computer vision algorithms. Here we have used a video camera to acquire live view of landslide site and a small computer board ‘raspberry Pi’ to run the algorithm we will introduce here. When landslide is detected SMS text messages are transmitted using a GSM modem. Due to video camera and Raspberry Pi, this method is inexpensive yet efficient and also requires low power, therefore this method can be used in any region. We have used median filtering to remove noise present in the detection algorithm. Here we have also proposed a new implementation method for median filtering using a linked list and parallel processing which is very time efficient.