AEROSPACE SHANGHAI, Volume. 42, Issue 2, 186(2025)
CNN-LSTM Based Space Object Recognition Method for Sequence Images
Figures & Tables(13)
Fig. 1. Architecture of a space object recognition network for sequential images
Fig. 2. Schematic diagram of the single frame feature extraction module design
Fig. 4. Schematic diagram of sampling the experimental sample sequence
Fig. 5. Confusion matrix of the validation set of the Top1 model from the noise experiment
Fig. 8. Confusion matrix of the validation set of the Top1 model from the attitude experiment
Fig. 9. Comparison of the activation maps for different feature extraction modules
Table 1.
Parameters for the network training View tableView in ArticleTable 1.
Parameters for the network training 训练轮数 学习率 权重衰减 动量 20 0.000 3 4×10-6 0.9
Table 2.
Comparison of the accuracy of noise experiment View tableView in ArticleTable 2.
Comparison of the accuracy of noise experiment 实验方法 Top1准确率/% mAP准确率/% 训练集 测试集 训练集 测试集 所提方法 99.63 99.44 99.77 99.48 VGG[ 31 ]+LSTM99.35 99.27 99.05 99.03 ResNet34[ 34 ]+LSTM98.10 83.92 93.15 65.52 EfficientNet-b0[ 35 ]+LSTM95.42 94.17 94.63 93.26 ViT16[ 36 ]+LSTM96.02 94.51 95.02 93.47 AlexNet[ 30 ]+Attention96.75 99.30 74.99 74.58
Table 3.
Comparison of the accuracy of randomly sampling attitudes View tableView in ArticleTable 3.
Comparison of the accuracy of randomly sampling attitudes 实验方法 Top1准确率/% mAP准确率/% 训练集 测试集 训练集 测试集 所提方法 96.19 94.27 94.36 93.53 VGG[ 31 ]+LSTM91.29 91.13 90.54 90.47 ResNet34[ 34 ]+LSTM91.05 56.62 90.57 55.19 EfficientNet-b0[ 35 ]+LSTM92.33 91.73 91.07 90.88 ViT16[ 36 ]+LSTM93.14 92.27 92.21 91.07 AlexNet[ 30 ]+Attention88.95 83.94 87.73 81.76
Table 4.
Comparisons of the recognition time and model size View tableView in ArticleTable 4.
Comparisons of the recognition time and model size 实验方法 识别时间/s 参数量（Flops）/M 所提方法 0.024 283.00 VGG[ 31 ]+LSTM0.047 6 011.30 ResNet34[ 34 ]+LSTM0.070 1 480.20 EfficientNet-b0[ 35 ]+LSTM0.032 435.16 ViT16[ 36 ]+LSTM0.099 17 630.20 AlexNet[ 30 ]+Attention0.042 2 731.60
Tools
Get Citation
Copy Citation Text
Siyu QI, Huijie ZHAO, Hongzhi JIANG, Xudong LI, Sihang WANG, Qi GUO. CNN-LSTM Based Space Object Recognition Method for Sequence Images[J]. AEROSPACE SHANGHAI, 2025, 42(2): 186
Paper Information
Category: Simulation and Analysis
Received: Jul. 15, 2024
Accepted: --
Published Online: May. 26, 2025
The Author Email:
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20