西安工业大学 光电工程学院,陕西 西安 710021
张尘(1998— ),女,硕士研究生,从事光场调控等方面的研究。1633825465@qq.com
lvhong@xatu.edu.cn
收稿:2025-03-13,
修回:2025-04-14,
纸质出版:2025-07-20
移动端阅览
张尘,吕宏,王鲲鹏等.基于深度学习与FPGA的涡旋光束OAM模式识别[J].光学技术,2025,51(04):392-399.
ZHANG Chen,LV Hong,WANG Kunpeng,et al.OAM mode recognition system of vortex beams based on deep learning and FPGA[J].OPTICAL TECHNIQUE,2025,51(04):392-399.
为提高结构化涡旋光束轨道角动量(OAM)的检测效率,提出结合深度学习与FPGA加速的OAM模式识别方法。利用功率谱反演法生成随机相位屏,通过空间光调制器模拟五种大气湍流环境,采集涡旋光束经过随机相位屏后的光斑数据图像,构建包含10000组样本的数据集,采用ResNet18卷积神经网络(CNN),对比分析不同湍流强度对涡旋光束OAM模式识别精度的影响,并对模型的识别性能进行全面评估,最终将训练后的ResNet18神经卷积网络模型载入可编程门阵列(FPGA)硬件平台实现OAM模式识别。结果表明:在弱、中湍流环境下的三种OAM模式识别准确率达100%,强湍流环境下的两种OAM模式识别准确率仍保持99%,相较于弱、中湍流环境下的OAM模式识别准确率略有下降。通过将CNN模型载入FPGA硬件平台,实现了耗能2.844/W、能效比2.19GOPS/W、推理时间0.58s的涡旋光束OAM模式识别。
To enhance the detection efficiency of structured vortex beams carrying Orbital Angular Momentum (OAM)
a recognition method combining deep learning with FPGA acceleration is proposed. Random phase screens are generated using the power spectrum inversion method
and five types of atmospheric turbulence conditions are simulated via a spatial light modulator. Vortex beam spot images after passing through the random phase screens are collected to construct a dataset comprising 10
000 samples. A ResNet18 Convolutional Neural Network (CNN) is employed to analyze and compare the impact of different turbulence intensities on OAM mode recognition accuracy. The recognition performance of the model is comprehensively evaluated
and the trained ResNet18 model is subsequently deployed onto a Field-Programmable Gate Array (FPGA) hardware platform for real-time OAM mode recognition. Experimental results demonstrate that under weak and moderate turbulence conditions
the recognition accuracy for three OAM modes reaches 100%
while under strong turbulence
the accuracy for two OAM modes remains as high as 99%
showing a slight decrease compared to the weaker turbulence cases. By deploying the CNN model onto the FPGA platform
the system achieves an energy consumption of 2.844 W
an energy efficiency of 2.19 GOPS/W
and an inference time of 0.58 seconds for OAM mode recognition of vortex beams.
娄岩 , 陈纯毅 , 赵义武 , 等 . 高斯涡旋光束在大气湍流传输中的特性研究 [J ] . 中国光学(中英文) , 2017 , 10 ( 6 ): 768 — 776 .
Lou Yan , Chen Chunyi , Zhao Yiwu , et al . Characteristics of gaussian vortex beam in atmospheric turbulence transmission [J ] . Chinese Optics , 2017 , 10 ( 6 ): 768 — 776 .
吕燕来 , 黄磊 , 高春清 . 光学坐标变换实现光束轨道角动量模式分束的技术研究 [J ] . 光学技术 , 2022 , 48 ( 06 ): 679 — 683 .
Lv Yanlai , Huang Lei , Gao Chunqing . Optical coordinate transformation-based technique for orbital angular momentum (OAM) mode beam splitting [J ] . Optical Technique , 2022 , 48 ( 06 ): 679 — 683 .
柯熙政 , 吴加丽 , 杨尚君 . 面向无线光通信的大气湍流研究进展与展望 [J ] . 电波科学学报 , 2021 , 36 ( 03 ): 323 — 339 .
Ke Xizheng , Wu Jiali , Shang jun . Research progress and prospects of atmospheric turbulence for wireless oftical communication [J ] Journd of Radio Science , 2021 , 36 ( 03 ): 323 — 339 .
Ming Z , Xindi T , Weixian W , et al . A novel deep-learning model compression based on filter-stripe group pruning and its IoT application [J ] . Sensors , 2022 , 22 ( 15 ): 5623 — 5623 .
吕浩 , 张盛兵 , 王佳 , 等 . 卷积神经网络SIP微系统实现 [J ] . 计算机工程与应用 , 2021 , 57 ( 05 ): 216 — 221 .
Lv Hao , Zhang Shengbing , Wang Jia , et al . Implementation of convolutional neural networks on a system-in-package (SiP) micro-system [J ] . Computer Engineering and Applications , 2021 , 57 ( 05 ): 216 — 221 .
Odetola T A , Groves K M , Mohammed Y , et al . 2L-3W: 2-Level 3-Way hardware-software co-verification for the mapping of convolutional neural network (CNN) onto FPGA boards [J ] . SN Computer Science , 2021 , 3 ( 1 ): 60 .
Lohani S , Knutson E M , O’Donnell M , et al . On the use of deep neural networks in optical communications [J ] . Applied Optics , 2018 , 57 ( 15 ): 4180 — 4190 .
Qingsong Z , Shiqi H , Yong W , et al . Mode detection of misaligned orbital angular momentum beams based on convolutional neural network [J ] . Applied Optics , 2018 , 57 ( 35 ): 10152 — 10158 .
Wang P , Zhang X , Fan D , et al . Convolutional neural network-assisted optical orbital angular momentum recognition and communication [J ] . IEEE Access , 2019 , 7 : 162025 — 162035 .
Xizheng K , Meng C . Recognition of orbital angular momentum vortex beam based on convolutional neural network [J ] . Microwave and Optical Technology Letters , 2021 , 63 ( 7 ): 1960 — 1964 .
郑崇辉 , 王天枢 , 刘哲绮 , 等 . 深度迁移学习方法识别轨道角动量光束 [J ] . 光电工程 , 2022 , 49 ( 06 ): 45 — 54 .
Zheng Conghui , Wang Tianshu , Liu Zheqi , et al . Recognition of orbital angular momentum beams using deep transfer learning methods [J ] . Opto-Electronic Engineering , 2022 , 49 ( 06 ): 45 — 54 .
Wang Z , Lai X , Huang H , et al . Recognizing the orbital angular momentum(OAM) of vortex beams from speckle patterns [J ] . Science China(Physics ,Mechanics & Astronomy), 2022 , 65 ( 04 ): 59 — 65 .
Zhang Z , Yin X , Cui X , et al . Performance analysis of modulating retro-reflector link based on orbital angular momentum coding in underwater channels [J ] . Optics Communications , 2022 , 510 : 127903 .
刘雪莲 , 陈旭东 , 林志立 , 等 . 深度学习辅助测量强散射涡旋光束拓扑荷数 [J ] . 光学学报 , 2022 , 42 ( 14 ): 209 — 215 .
Liu Xuelian , Chen Xundong , Lin Zhili , et al . Deep learning-assisted measurement of topological charge in strongly scattered vortex beams [J ] . Acta Optica Sinica , 2022 , 42 ( 14 ): 209 — 215 .
M Z M K , A M R , M M , et al . L-band InAs/InP quantum dash laser spatial OAM light modes classification under smoke environment: an image processing enhanced deep learning approach [J ] . Optics and Laser Technology , 2024 , 168 : 109933 .
孙立虎 , 李萍萍 , 刘素娟 , 等 . 基于矢量涡旋光束和深度学习的图像信息编解码 [J ] . 激光与光电子学进展 , 2025 , 62 ( 10 ): 1007001 .
Lihu Sun , Pingping Li , Sujuan Liu , et al . Image information encoding and decoding based on vector vortex beam and deep learning [J ] . Laser & Optoelectronics Progress , 2025 , 62 ( 10 ): 1007001 .
0
浏览量
1
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010802024621