Research on Video Detection Method of Mudslide based on Inflated 3D Convolutional Neural Network

Zefeng Yu

doi:10.54097/fcis.v5i1.12004

Authors

Zefeng Yu

DOI:

https://doi.org/10.54097/fcis.v5i1.12004

Keywords:

Debris Flow Transformation, Weak Labeling, Convolutional Networks

Abstract

Mudslide is a common natural disaster in mountainous areas, causing harm to roads and railroads and structures. To address this problem, this paper adopts the automatic video recognition approach, which utilizes widely installed video surveillance equipment to detect the changes of mudslides, so as to identify the mudslide diffuse flow disaster in the video monitoring area to achieve early warning. Firstly, the deep learning model is trained with weakly labeled mudslide video files, and the spatio-temporal feature learning method, i.e., inflated 3D convolutional network, is combined in the model, which results in a higher correctness rate of training and detection; secondly, the model is shown to have a recognition accuracy of 86% through the testing of the relevant datasets, which can be used as an effective complement to the traditional method of mudslide early warning.

Downloads

Download data is not yet available.

References

Lohumi K, Roy S. Automatic detection of flood severity level from flood videos using deep learning models[C]//2018 5th International Conference on Information and Communication Technologies for Disaster Management (ICT-DM). IEEE, 2018: 1-7.

Lopez-Fuentes L, van de Weijer J, Bolanos M, et al. Multi-modal Deep Learning Approach for Flood Detection[J]. MediaEval, 2017, 17: 13-15.

Pang G, Shen C, Cao L, et al. Deep learning for anomaly detection: A review[J]. ACM computing surveys (CSUR), 2021, 54(2): 1-38.

Soltani Nejad S. Weakly-Supervised Anomaly Detection in Surveillance Videos Based on Two-Stream I3D Convolution Network[J]. 2023.

Munukutla P S, Jain S. One Shot Learning for Video Object Segmentation using Fully Convolutional I3D Network[J].

Kexuan W, Yifei X, Xinyu X. Research on Machine Learning Based on Multi-label Algorithm[C]//2020 IEEE International Conference on Artificial Intelligence and Computer Applications (ICAICA). IEEE, 2020: 824-829.

Huaihui C, Shengbo Z. Multi Instance Deep Learning Target Tracking[J]. The Frontiers of Society, Science and Technology, 2020, 2(9).

Chaganti P C V, Vasireddy K, Reddy E R, et al. Predicting Landslides and Floods with Deep Learning[C]//2023 4th International Conference on Electronics and Sustainable Communication Systems (ICESC). IEEE, 2023: 1259-1265.

Li L, Doroslovački M, Loew M H. Approximating the gradient of cross-entropy loss function[J]. IEEE access, 2020, 8: 111626-111635.

Tran D, Bourdev L, Fergus R, et al. Learning spatiotemporal features with 3d convolutional networks[C]//Proceedings of the IEEE international conference on computer vision. 2015: 4489-4497.