Deep learning based optical curvature sensor through specklegram detection of multimode fiber

Li, Guangde; Liu, Yan; Qin, Qi; Zou, Xiaoli; Wang, Muguang; Yan, Fengping

doi:10.1016/j.optlastec.2022.107873

Cited by 40 publications

(13 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although the accuracy obtained in our work is lower in comparison with the models proposed in Refs. 37, 39, 42, and 43, the model is still able to successfully recognize with a relatively fewer number of classes (seven weights) with a considerable accuracy of 93.1%. This is still higher than that presented in Refs.…”

Section: Resultsmentioning

confidence: 95%

“…32,33 Furthermore, a paradigm shift has been observed toward the development of neural networks-assisted FSSs in the recent past for various purposes. [34][35][36][37][38][39][40][41][42][43] Rodríguez-Cuevas et al proposed a convolutional neural network (CNN) model with a classification accuracy of 99% for three locations and 79% for ten locations of vibrations. 34 Razmyar and Mostafavi 35 estimated and classified the deflection direction of the MMF and achieved 96% classification accuracy.…”

Section: Introductionmentioning

confidence: 99%

“…Furthermore, a paradigm shift has been observed toward the development of neural networks-assisted FSSs in the recent past for various purposes 34 – 43 Rodríguez-Cuevas et al. proposed a convolutional neural network (CNN) model with a classification accuracy of 99% for three locations and 79% for ten locations of vibrations 34 .…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Analyzing specklegrams of plastic optical fiber using convolutional neural network for weight recognition

et al. 2022

View full text Add to dashboard Cite

A weight recognition scheme is proposed based on the analysis of plastic optical fiber (POF) specklegrams. A simple experiment is performed for data acquisition comprising specklegram images corresponding to different external applied weights ranging from 0 to 3 kg in steps of 0.5 kg on the POF. These specklegram images are further split into training, validation, and test datasets for employment in the convolutional neural network (CNN). The model is trained and validated with a sufficient number of epochs (iterations) to obtain optimal training and validation accuracies (>90%); it is then used for recognizing (classifying) the unseen test dataset images. A user-defined CNN model is optimized using four different optimizers: Adam, AdaMax, Nadam, and RMSProp. The recognition (or test) accuracy of these optimizers is compared. The Nadam optimizer has the highest recognition accuracy of 93.1% for increasing weights and 91.9% for decreasing weights. Furthermore, we investigate the impact of two parameters, namely temperature effect and information loss of specklegrams, on the model's weight recognition accuracy. The temperature effect is studied for finite temperatures ranging from 29°C to 35°C with 3°C step fluctuations around the ambient temperature of 25°C. We find that the maximum deviation in recognition accuracy is about 1.2%. To quantify the information content (speckles) in each blocked specklegram, Shannon entropy (SE) is estimated. These images are then used in our existing CNN model with the Nadam optimizer to evaluate recognition performance with blocking. With increasing blocking, we observe a decrease in SE and recognition accuracy. This analysis suggests that, even with information loss in the specklegrams at ≤20% blocking, a good representation of weights with high recognition accuracy (>80%) can be obtained. The presence of a small physical obstacle in front of the imaging system can cause information loss in specklegrams to manifest in practical scenarios.

show abstract

Section: Resultsmentioning

confidence: 95%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Analyzing specklegrams of plastic optical fiber using convolutional neural network for weight recognition

et al. 2022

View full text Add to dashboard Cite

show abstract

“…However, most of these methods analyze the sensing parameters by calculating the correlation coefficients between the speckle patterns, resulting in limited measurement range and resolution. Deep learning-based demodulation schemes have also been proposed and proven effective through rigorous experiments [ 21 , 22 , 23 , 24 , 25 , 26 ]. Specifically, the speckle patterns collected from the fiber output and the corresponding sensing parameters are considered training samples and labels of the dataset, respectively, and the produced dataset is used to train the neural network.…”

Section: Introductionmentioning

confidence: 99%

Demonstration of a Learning-Empowered Fiber Specklegram Sensor Based on Focused Ion Beam Milling for Refractive Index Sensing

Gao

2023

Nanomaterials

View full text Add to dashboard Cite

We report a simple and robust fiber specklegram refractive index sensor with a multimode fiber-single mode fiber-multimode fiber structure based on focused ion beam milling. In this work, a series of fluid channels are etched on the single-mode fiber by using focused ion beam milling to enhance the interaction between light and matter, and a deep learning model is employed to demodulate the sensing signal according to the speckle patterns collected from the output end of the multimode fiber. The feasibility and effectiveness of the proposed scheme were verified by rigorous experiments, and the test results showed that the demodulation accuracy and speed could reach 99.68% and 4.5 ms per frame, respectively, for the refractive index range of 1.3326 to 1.3679. The proposed sensing scheme has the advantages of low cost, easy implementation, and a simple measurement system, and it is expected to find applications in various chemical and biological sensing.

show abstract

“…Deep learning has made remarkable achievements in computer science and engineering-related fields [12][13][14][15][16][17][18], demonstrating the potential of deep learning to overcome the shortcomings of traditional methods and create unprecedented opportunities in these fields [19]. Inspired by the above success, researchers have applied deep learning in the field of fiber sensing [20][21][22][23][24], using neural networks to analyze speckle patterns. The learning-based fiber specklegram sensor has high accuracy, and its measurement range only depends on the calibration range.…”

Section: Introductionmentioning

confidence: 99%

Bending recognition based on learning enhanced fiber specklegram sensor

Gao,

Zhang,

Zhang

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

In this paper, we propose and demonstrate a deep learning-enhanced fiber specklegram sensor for bending recognition. A segment of multimode fiber is used to sense bending, and tiny bending change leads to significant variations in the speckle pattern generated at the distal end of the fiber. Principal component analysis is utilized to optimize the collected samples and remove noise and redundant information by mining internal features, which makes the mapping relationship between speckle image and corresponding curvature clearer and is conducive to reducing computational complexity. BP neural network is employed to learn the mapping relationship between speckle image and curvature on the optimized dataset according to the optimized direction provided by principal component analysis. The testing results show that the prediction error of the trained model for the learned bending state is 5.910-4 m-1, and the prediction speed is 0.05 milliseconds per frame. The proposed scheme has a strong generalization ability and can be applied to predict bending states that have never been learned or seen with a prediction error of 3.810-2 m-1, which cannot be realized by the previously reported fiber specklegram sensor based on the classification neural network. The bending recognition scheme enhanced by deep learning proposed in this paper provides an enlightening reference for solving fiber sensing problems with deep learning methods, and it has the potential to be applied in more fields as a general scheme.

show abstract

Deep learning based optical curvature sensor through specklegram detection of multimode fiber

Cited by 40 publications

References 27 publications

Analyzing specklegrams of plastic optical fiber using convolutional neural network for weight recognition

Analyzing specklegrams of plastic optical fiber using convolutional neural network for weight recognition

Demonstration of a Learning-Empowered Fiber Specklegram Sensor Based on Focused Ion Beam Milling for Refractive Index Sensing

Bending recognition based on learning enhanced fiber specklegram sensor

Contact Info

Product

Resources

About