基于改进VisionTransformer的森林火灾视频识别研究

打开文本图片集

键词：森林火灾；深度学习；目标检测；三维卷积神经网络；VisionTransformer中图分类号：S762 文献标志码：A 文章编号：1000-2006（2025）04-0186-09

Research on forest fire video recognition based on improved Vision Transformer

ZHANG Min，XIN Ying*，HUANG Tianqi

（College of Mechanical and Electrical Engineering，Northeast Forestry University，Harbin15O04O，China）

Abstract：【Objective】Thisresearchaimstoresolvethelimitationsofexistingforestfirerecognitionalgorithmsin temporal featureutilizationandcomputational eficiency，this studyproposesavideo-basedrecognitionmodel（C3DViT）to enhance bothdetection acuracyandoperational eficiencyin practical forest monitoring scenarios.【Method】We presentedahybridarchitectureintegrating 3DConvolutional Neural Networks（3DCNN）with Vision Transformer（ViT）. Theframework emploied3Dconvolution kernels to extract spatiotemporal features from video sequences，which were subsequently tokenized intovectorrepresentations.Vision Transformer'sself-atentionmechanism thengloballymodels feature relationshipsacross temporalandspatial dimensions，with final classificationachievedthroughtheMLPHead layer.Comprehensive ablation studiedandcomparative experiments were conductedagainst ResNet5O，LSTM，YOLOv5， and baseline 3DCNN，ViT models.【Result】The C3D-ViT achieves 96. 10% accuracy，outperforming ResNet50 89.07% ），LSTM 93.26% ），and YOLOv5（ 91.46% ），and hasimproved compared to the accuracyof the original 3DCNN and Vision Transformer（ 93.91% ， 90.43% ）.The improved C3D-ViT model performs better in recognition performance，with highrecognition accuracyand stabilityunder unfavorable conditions such asocclusion，long distance， andthinsmoke.Thedemandforreal-timedetectioncanberealized.【Conclusion】The C3D-ViTframework ffectively addresses spatiotemporal modelingchallngesinwildfiredetection through synergistic CNN-Transformerinteraction， providing a technically viable solution for next-generation forest fire early warning systems.

Keywords：forest fire；deep learning；object detection;3DCNN;Vision Transformer（ViT）

森林是陆地生态系统重要的组成部分，在全球生态、碳循环及气候演变中起着重要作用[1]。（剩余16718字）