Research Article
Research on Moving Target Behavior Recognition Method Based on Deep Convolutional Neural Network
@INPROCEEDINGS{10.1007/978-3-030-67874-6_28, author={Jian-fang Liu and Hao Zheng and He Peng}, title={Research on Moving Target Behavior Recognition Method Based on Deep Convolutional Neural Network}, proceedings={Advanced Hybrid Information Processing. 4th EAI International Conference, ADHIP 2020, Binzhou, China, September 26-27, 2020, Proceedings, Part II}, proceedings_a={ADHIP PART 2}, year={2021}, month={1}, keywords={Convolutional neural network Moving target Recognition Depth}, doi={10.1007/978-3-030-67874-6_28} }- Jian-fang Liu
Hao Zheng
He Peng
Year: 2021
Research on Moving Target Behavior Recognition Method Based on Deep Convolutional Neural Network
ADHIP PART 2
Springer
DOI: 10.1007/978-3-030-67874-6_28
Abstract
In order to solve the problem that the average recognition degree of moving target line is low by the traditional method of moving target behavior recognition. Therefore, a motion recognition method based on deep convolutional neural network is proposed. Construct a deep convolutional neural network target model, and use the model to design the basic unit of the network. The returned unit is calculated to the standard density map by the set unit, and the moving target position is determined by the local maximum method to realize the moving target behavior recognition. The experimental results show that The experimental results of the multi-parameter SICNN256 model are slightly better than other model structures. And the average recognition rate and the recognition rate of the moving target behavior recognition method based on deep convolutional neural network are higher than the traditional method, which proves its effectiveness. Since a single target is more frequent than multiple recognitions and there is no target similar recognition, similar target error detection cannot be excluded.
