Journals >Optics and Precision Engineering
Contents
2023
Volume: 31 Issue 18
13 Article(s)
Export citation format
Information Sciences
Global and local feature fusion image dehazing
Xin JIANG, Haitao NIE, and Ming ZHU
Convolution operations with parameter sharing features primarily focus on the extraction of local features of images but fail to model the features beyond the range of the receptive field. Moreover， when the parameters of an entire image share the same convolution kernel， the characteristics of different regions are igConvolution operations with parameter sharing features primarily focus on the extraction of local features of images but fail to model the features beyond the range of the receptive field. Moreover， when the parameters of an entire image share the same convolution kernel， the characteristics of different regions are ignored. To address this limitation in existing methods， a global and local feature fusion dehazing network is proposed. We utilize transformer and convolution operations to extract global and local feature information from images， respectively. Subsequently， we merge and output these features， effectively employing the advantages of transformers in modeling long-distance dependencies and the local perception of convolution operations， thus achieving efficient feature expression. Before the final output of restored images， we incorporate an enhancement module that includes multi-scale patches to further aggregate global feature information and enhance the details of the restored images using a transformer. Simultaneously， we introduce a global positional encoding generator， which can adaptively generate positional encodings based on the global content information of images， thereby enabling 2D spatial location modeling of the dependency relationship between pixels. Experimental results demonstrate the superior performance of the proposed dehazing network on both synthetic and real image datasets， producing more realistic restored images and significantly reducing detail loss..
Optics and Precision Engineering
- Publication Date: Sep. 25, 2023
- Vol. 31, Issue 18, 2687 (2023)
Cross-scale and cross-dimensional adaptive transformer network for colorectal polyp segmentation
Liming LIANG, Anjun HE, Renjie LI, and Jian WU
To address the problem of large-scale variation， blurred boundaries， irregular shapes， and low contrast with normal tissues in colon polyp images， which leads to the loss of edge detail information and mis-segmentation of lesion areas， we propose a cross-dimensional and cross-scale adaptive transformer segmentation netTo address the problem of large-scale variation， blurred boundaries， irregular shapes， and low contrast with normal tissues in colon polyp images， which leads to the loss of edge detail information and mis-segmentation of lesion areas， we propose a cross-dimensional and cross-scale adaptive transformer segmentation network. First， the network uses transformer encoders to model the global contextual information of the input image and analyze the colon polyp lesion areas at multiple scales. Second， the channel attention and spatial attention bridges are used to reduce channel dimension redundancy and enhance the model's spatial perception ability while suppressing background noise. Third， the multi-scale dense parallel decoding module is used to bridge the semantic gaps between cross-scale feature information at different layers， effectively aggregating multi-scale contextual features. Fourth， a multi-scale prediction module is designed for edge details， guiding the network to correct boundary errors in a learnable manner. The experimental results conducted on the CVC-ClinicDB， Kvasir-SEG， CVC-ColonDB， and ETIS datasets showed that the Dice similarity coefficients are 0.942， 0.932， 0.811， and 0.805， and the average intersection-over-union ratios are 0.896， 0.883， 0.731， and 0.729， respectively. The segmentation performance of our proposed method was better than that of existing methods. The simulation experiment showed that our method can effectively improve the mis-segmentation of colon polyp lesion areas and achieve high segmentation accuracy， providing a new approach for colon polyp diagnosis..
Optics and Precision Engineering
- Publication Date: Sep. 25, 2023
- Vol. 31, Issue 18, 2700 (2023)
Uniform defocus blind deblurring based on deeper feature-based wiener deconvolution
Chengxi WANG, Chen LUO, Jianghao ZHOU, Lang ZOU, and Lei JIA
In industrial precision manufacturing， the small field depths of the imaging systems of visual inspection equipment can make them susceptible to defocus blurring. This significantly degrades their detection effect. To address this issue， this paper proposes a uniform defocus blind deblurring network （UDBD-Net）. First， In industrial precision manufacturing， the small field depths of the imaging systems of visual inspection equipment can make them susceptible to defocus blurring. This significantly degrades their detection effect. To address this issue， this paper proposes a uniform defocus blind deblurring network （UDBD-Net）. First， a uniform defocus blur kernel estimation net for extracting the characteristics of out-of-focus blurring and accurately estimating the blur kernel is proposed. Second， a non-blind deconvolution network， which is used for learning and estimating the unknown quantity in the feature-based Wiener deconvolution （FWD） formula so as to accurately generate the latent features of blurred images， is presented. Finally， the use of an encoder–decoder net to enhance the details of the recovered image and remove the artifacts is detailed. The experimental results indicate peak signal-to-noise ratio （PSNR） values of 31.16 dB and 36.16 dB for UDBD-Net on the images of DIV2K and GOPRO test sets， respectively. Compared with extant blind deblurring methods， the proposed method can restore deblurred images with higher quality and more naturalness without significantly increasing the model inference time. Furthermore， UDBD-Net can achieve a good deblurring effect on real uniformly defocused blurred images and can considerably improve the detection effect of industrial vision detection algorithms on such images..
Optics and Precision Engineering
- Publication Date: Sep. 25, 2023
- Vol. 31, Issue 18, 2713 (2023)
Joint self-attention and branch sampling for object detection on drone imagery
Yunzuo ZHANG, Cunyu WU, Yameng LIU, Tian ZHANG, and Yuxin ZHENG
Object detection on drone imagery is widely used in many fields. However， due to the complexity of the image background， the dense small objects and the dramatic scale changes， the existing object detection on drone imagery methods are not accurate enough. In order to solve this problem， we propose an accurate object dObject detection on drone imagery is widely used in many fields. However， due to the complexity of the image background， the dense small objects and the dramatic scale changes， the existing object detection on drone imagery methods are not accurate enough. In order to solve this problem， we propose an accurate object detection method for drone imagery joint self attention and branch sampling. Firstly， a nested residual structure integrating self attention and convolution is designed to achieve the effective combination of global and local information， which makes the model to focus on the object area and ignore invalid features. Secondly， we design a feature fusion module based on branch sampling to mitigate the loss of object information. Finally， an improved detector for small objects is added to alleviate the problem of sharp scale changes. Furthermore， we propose a feature enhancement module to obtain more discriminative small object features. The experimental results show that the proposed algorithm performs well in various scenarios. Specifically， the mAP50 and mAP of the s model on the VisDrone2019 reached 59.3% and 37.1% respectively， an increase of 5.6% and 5.4% compared with the baseline. The mAP50 and mAP on the UAVDT reached 44.1% and 24.9% respectively， an increase of 5.8% and 3.2% compared with the baseline..
Optics and Precision Engineering
- Publication Date: Sep. 25, 2023
- Vol. 31, Issue 18, 2723 (2023)
Weakly supervised video instance segmentation with scale adaptive generation regulation
Yinhui ZHANG, Weiqi HAI, Zifen HE, Ying HUANG, and Dongdong CHEN
Video instance segmentation is critical in multi-target perception and scene understanding in assisted driving. However， as weakly supervised video instance segmentation is often applied to bounding box annotations for network training， the segmentation accuracies of targets with large-scale dynamic ranges in traffic sVideo instance segmentation is critical in multi-target perception and scene understanding in assisted driving. However， as weakly supervised video instance segmentation is often applied to bounding box annotations for network training， the segmentation accuracies of targets with large-scale dynamic ranges in traffic scenes are severely restricted. To address this issue， we propose a scale adaptive generation regulation weakly supervised video instance segmentation network （SAGRNet）. First， a multi-scale feature mapping contribution dynamic adaptive control module is proposed to replace the original linear weighting. This enables placing the focus on the local position and global contour of the target by dynamically adjusting the contribution of different scale feature mapping information， which solves the problem of large-scale dynamic ranges caused by changes in the imaging distance between vehicles and pedestrians. Second， a target instance multi-fine-grained spatial information aggregation generation control module is constructed to regulate the feature maps of each scale using weight parameters， which are obtained by aggregating multi-fine-grained spatial information extracted based on different dilations. This module refines the instance boundary and improves the representation of cross-channel mask interaction information， effectively compensating for the lack of edge contour segmentation mask continuity caused by limited instance edge information. Finally， to alleviate the weak supervision derived from bounding box level annotations， orthogonal and color similarity losses are introduced to reduce the deviation between the model prediction mask and real bounding box and to address the pixel-wise label attribute classification ambiguity problem. Experimental results on a traffic scene dataset extracted from Youtube-VIS2019 indicate that the SAGRNet improves the mean accuracy by 5.1% to 38.1% compared with the weakly supervised baseline. These results prove that our method provides an effective theoretical basis for multi-target perception and instance level scene understanding..
Optics and Precision Engineering
- Publication Date: Sep. 25, 2023
- Vol. 31, Issue 18, 2736 (2023)
Spatial-spectral Transformer for classification of medical hyperspectral images
Yuan LI, Xu SHI, Zhengchun YANG, Qijuan TAN, and Hong HUANG
The development of hyperspectral imaging （HSI） technology offers new avenues for non-invasive medical imaging. However， medical hyperspectral images are characterized by high dimensionality， high redundancy， and the property of “graph-spectral uniformity，” necessitating the design of high-precision diagnostic algorithmThe development of hyperspectral imaging （HSI） technology offers new avenues for non-invasive medical imaging. However， medical hyperspectral images are characterized by high dimensionality， high redundancy， and the property of “graph-spectral uniformity，” necessitating the design of high-precision diagnostic algorithms. In recent years， transformer modes have been widely applied in medical hyperspectral image processing. However， medical hyperspectral images obtained using various instruments and acquisition methods have significant differences； this considerably hinders the practical applications of existing transformer-based diagnostic models. To address the aforementioned issues， a spatial–spectral self-attention transformer （S3AT） algorithm is proposed to adaptively mine the intrinsic relations between pixels and bands. First， in the transformer encoder， a spatial–spectral self-attention mechanism， which is designed to obtain key spatial information and important bands on hyperspectral images from different viewpoints， is employed. Thus， the spectral–spectral self-attention obtained from different views is fused. Second， in the classification stage， the predictions from different views are fused according to the learned weights. The experimental result on in-vivo human brain and blood cell HSI datasets indicate that the overall classification accuracies reach 82.25% and 91.74%， respectively. This demonstrates that the proposed S3AT algorithm yields enhanced classification performance on medical hyperspectral images..
Optics and Precision Engineering
- Publication Date: Sep. 25, 2023
- Vol. 31, Issue 18, 2752 (2023)
Polyp image segmentation based on multi-scale ResNeSt-50 aggregation network and message passing
Ping XIA, Guangyi ZHANG, Bangjun LEI, Yaobing ZOU, and Tinglong TANG
There boundary between colorectal polyps and normal tissues is not typically evident. Therefore， accurately locating polyp positions is challenging. This study developed a novel polyp image segmentation method based on a combination of multiscale ResNeSt-50 aggregation network and sequential tree-reweighted message pasThere boundary between colorectal polyps and normal tissues is not typically evident. Therefore， accurately locating polyp positions is challenging. This study developed a novel polyp image segmentation method based on a combination of multiscale ResNeSt-50 aggregation network and sequential tree-reweighted message passing （TRW-S）. First， a multiscale ResNeSt-50 aggregation network with an encoding–decoding structure was constructed to improve the expressiveness of the network. The encoder of the network is cascaded by convolution module and four-level ResNeSt module to build the ResNeSt-50 backbone network， which realizes linear integration and communication between cross-channel information， ResNeSt-50 uses split attention to strengthen the performance of important channel groups and enhance the ability of the residual module to extract polyp image information. In the bottom three layers of the decoder， a multilayer receptive field block （RFB） was used to obtain multiscale information. Subsequently， the dense aggregation module was used to integrate the output. The decoding information was output by using a fast decoding method， which ensured consistent segmentation performance and reduced the number of parameters. Second， the test-time augmentation （TTA） module was used to improve the prediction accuracy and enhance the generalization ability of the network when generating predictive images. Finally， a sequential tree-reweighted message passing （TRW-S） algorithm based on Markov random fields was constructed to postprocess the predicted image output of the model. This helped achieve continuity of the segmentation edge and consistency within the segmentation region. The experimental results on Kvasir-SEG， an open-access dataset for gastrointestinal polyps images， show that our method achieved an mDice value of 91.6%， mIoU of 86.3%， Smeasure of 92.1%， and MAE of 2.3%，which are higher than those of the polyp segmentation algorithms based on U-NET， U-Net++， ResUNet， SFA， and PraNet. Test results on the unknown datasets ETIS-LaribPolypDB and ColonDB indicate that the proposed model affords improvements in the PraNet and mDice values by 16.4% and 7.7%， respectively. As regards the segmentation performance on the ETIS-LaribPolypDB dataset， the proposed model was found to be highly sensitive to small lesions. Thus， the proposed model exhibits excellent performance in terms of consistency of segmentation area， continuity of segmentation edge， sharpness of contour， and ability to capture small lesions. In addition， it exhibits good generalization ability in the case of unknown datasets..
Optics and Precision Engineering
- Publication Date: Sep. 25, 2023
- Vol. 31, Issue 18, 2765 (2023)
Micro/Nano Technology and Fine Mechanics
Sensitivity laws and prediction of piezoelectric force sensors at different loading points
Jun ZHANG, Tiantian ZHEN, Jiale CAI, Mengtong LI, and Yuting LIU
A multi-point supported piezoelectric force sensor was designed to measure a wide range of vector forces with variable application points. More specifically， this study focused on analyzing and predicting sensitivity variations of the force sensor at various application points to ensure precise measurements of vector fA multi-point supported piezoelectric force sensor was designed to measure a wide range of vector forces with variable application points. More specifically， this study focused on analyzing and predicting sensitivity variations of the force sensor at various application points to ensure precise measurements of vector forces at different loading points. The factors that cause sensitivity variations of the force sensor at different test positions were analyzed， and the relationship between the sensitivity at such points within the working surface of the force sensor and the force-electricity conversion coefficient of the force measuring unit was derived. A least-squares support vector machine （LS-SVM） prediction model was established. Verification experiments showed that this model accurately predicts sensitivity values for the force sensor at different application points， with an error below 3% compared to actual values. The LS-SVM model offers the advantages of speed， reliability， and high precision in predicting sensitivity for multi-point supported piezoelectric force sensors at various application points， indicating its efficacy in quantitatively analyzing complex variable relationships..
Optics and Precision Engineering
- Publication Date: Sep. 25, 2023
- Vol. 31, Issue 18, 2656 (2023)
Nonlinear decoupling of 3d force sensors based on WSO-ELM
Shizheng SUN, Ke PANG, Jingtong YU, and Renxiang CHEN
A nonlinear decoupling algorithm based on the white shark optimization algorithm optimized extreme learning machine （WSO-ELM） is proposed to address the issue of inter dimensional coupling interference in three-dimensional force sensors， with an integrated three-dimensional force sensor based on fiber Bragg grating （FBA nonlinear decoupling algorithm based on the white shark optimization algorithm optimized extreme learning machine （WSO-ELM） is proposed to address the issue of inter dimensional coupling interference in three-dimensional force sensors， with an integrated three-dimensional force sensor based on fiber Bragg grating （FBG） as the research object. Firstly， an integrated three-dimensional force sensor based on FBG was designed to reveal the mapping relationship between the wavelength drift of the sensor and the three-dimensional force； Then， a static calibration experimental system is established to analyze the three-dimensional force coupling characteristics， and a WSO-ELM algorithm three-dimensional force sensor decoupling model is established. The model is optimized using the stable and efficient characteristics of the white shark optimization algorithm （WSO） to find the optimal parameter combination of the number of neurons in the hidden layer of the ELM neural network and the decoupling time. Research on nonlinear decoupling of three-dimensional force sensors based on WSO-ELM is carried out； Finally， after decoupling， the maximum average type I error of the sensor reaches 0.51%， and the maximum average type II error reaches 0.65%， achieving three-dimensional force nonlinear decoupling based on WSO-ELM. To verify the decoupling effect， a comparative experiment was conducted between the WSO-ELM algorithm and the extreme learning machine neural network model， backpropagation neural network， and least squares method for decoupling effect. The experimental results show that the WSO-ELM algorithm has good decoupling effect， can effectively construct the coupling relationship between three-dimensional force dimensions， reduce sensor coupling interference， improve sensor measurement accuracy， and has good nonlinear decoupling ability..
Optics and Precision Engineering
- Publication Date: Sep. 25, 2023
- Vol. 31, Issue 18, 2664 (2023)
Full leaf-spring type spatial large-stroke parallel flexure micro-positioning stage and trajectory control
Yunzhuang CHEN, Leijie LAI, Pengzhi LI, and Limin ZHU
To address the drawbacks of the traditional micropositioning stage， such as the small range of motion， low motion accuracy caused by parasitic motion， and serious cross-axis coupling， this paper proposes a full leaf-spring parallel flexure decoupling micropositioning stage driven by a voice coil motor with large-strokeTo address the drawbacks of the traditional micropositioning stage， such as the small range of motion， low motion accuracy caused by parasitic motion， and serious cross-axis coupling， this paper proposes a full leaf-spring parallel flexure decoupling micropositioning stage driven by a voice coil motor with large-stroke and multiple degrees of freedom （multi-DOF）. First， the structure and deformation principle of the long-stroke multi-DOF parallel flexure mechanism with a leaf-spring type flexure spherical joint are introduced. Second， considering 3-DOF as an example， the kinematic equation of the moving platform is derived， the input stiffness model of the mechanism is established， and the compliance modeling and design of the flexure spherical joint are provided based on the compliance matrix method to determine the parameters of the micropositioning stage. Additionally， the models of the system dynamics are identified for 3-DOF. On the basis of the models， a composite controller of phase advanced proportional-integral （PI） feedback control combined with sliding mode feedforward control is designed. Finally， a stage experimental system is developed to verify its trajectory tracking performance. Experimental results indicate that， compared with the classical proportional-integral-derivative （PID） control， the compound control method can improve the track tracking performance by more than 95% and that the added sliding mode feedforward effectively eliminates the phase lag caused by simple feedback control. Meanwhile， the proposed multi-DOF micropositioning stage can achieve a motion with ±3.23 mm×±21.50 mrad×±20.30 mrad. It has the characteristics of large stroke， good stability， and high accuracy， which are applicable in many spatial positioning situations that require large travel and high accuracy..
Optics and Precision Engineering
- Publication Date: Sep. 25, 2023
- Vol. 31, Issue 18, 2675 (2023)
Modern Applied Optics
Effect of magnetic field and target-substrate distance on properties of CrAlN films
Jinhua ZHENG, Shiyang MEI, Zhixiong LI, and Qingyun LIU
To improve the binding strength between a chromium aluminum nitride （CrAlN） film and substrate and enhance the film's wear resistance， a CrAlN film was deposited on a 45 steel substrate via cathodic arc sputtering. The influence mechanisms of a magnetic field and the target-substrate distance on the properties of tTo improve the binding strength between a chromium aluminum nitride （CrAlN） film and substrate and enhance the film's wear resistance， a CrAlN film was deposited on a 45 steel substrate via cathodic arc sputtering. The influence mechanisms of a magnetic field and the target-substrate distance on the properties of the CrAlN film were analyzed. In addition， the effects of the magnetic field on the film's thickness， surface morphology， binding force， and wear resistance were investigated and compared with those for a film deposited without a magnetic field. The results revealed that the film thickness， surface roughness， droplet size， and number of droplets decreased with an increase in the target-substrate distance. Compared with the sample deposited without a magnetic field， the surface roughness values of the samples deposited with a magnetic field were smaller， the radii of curvature of the films were larger， and the binding forces were better. The interfacial binding force was approximately 60 N with the magnetic field， and it changed little with the target-substrate distance， whereas the binding force changed significantly without a magnetic field. When the two types of samples were subjected to the same conditions， the binding force increased by 20%-80%. The friction coefficient and wear rate of the film decreased with an increase in the target-substrate distance， and the friction coefficient and wear rate of the sample with a magnetic field were smaller at the same target-substrate distance. When a magnetic field was applied and the target-substrate distance was 160-180 mm， the deposited CrAlN film exhibited the best performance. These results provide important references for the preparation of high-performance CrAlN films..
Optics and Precision Engineering
- Publication Date: Sep. 25, 2023
- Vol. 31, Issue 18, 2627 (2023)
Study on the characteristics of four electrode wide constant temperature region for preparing high quality spiral fiber
Jian WANG, Chao MA, Donghui WANG, Hongye WANG, and Libo YUAN
To prepare high-quality helical fiber devices， a heating method for an arc fiber with four electrodes was developed. A finite-element simulation model of arc plasma discharge with four electrodes was established based on the magnetohydrodynamics model. The effects of the electrode angle and distance and applied voltageTo prepare high-quality helical fiber devices， a heating method for an arc fiber with four electrodes was developed. A finite-element simulation model of arc plasma discharge with four electrodes was established based on the magnetohydrodynamics model. The effects of the electrode angle and distance and applied voltage on the temperature field of arc discharge were studied. A temperature-field simulation diagram was created based on the actual situation， and it indicated that the electrode rod with four electrodes formed a wide constant-temperature region. The wide constant-temperature zone was conducive to the release of fiber stress and reduced the influence of fiber migration. Based on the simulation structure， a plasma hot melt torsional-processing system with a wide constant-temperature zone for arc discharge with four electrodes was developed. When the maximum temperature of the heated fiber in the system was approximately 1 050 ℃， the axial constant-temperature zone length of the heated fiber was approximately 2.12 mm. Finally， helical long-period fiber gratings with different optical fibers and different periods were prepared by using the developed system. The experimental results revealed that when the wavelength of the prepared helical long-period fiber gratings was 1.21~1.3 µm， the intensity fluctuation of the transmission spectrum was less than 1 dB and the average intensity was higher than -1 dB. At wavelengths of 1.3~1.35 µm， the average light intensity was higher than -1 dB. The deepest trough of the transmission spectrum was >22 dB. To further verify the performance of the system， helical fiber structures with different periods were prepared using a single-mode， eccentric core optical， and eccentric twin-core fibers. The results showed that the cladding boundary of the fabricated helical fiber structure was clear and straight and no significant thread structure was present. No distinctive helical machining marks were observed in the central core of the fiber. The off-core of the fiber was smooth and continuous..
Optics and Precision Engineering
- Publication Date: Sep. 25, 2023
- Vol. 31, Issue 18, 2636 (2023)
Measurement of two-dimensional distribution of stress birefringence based on dual photoelastic modulators cascade difference frequency modulation
Kewu LI, Yuanfang QIU, Zhiying CUI, Lei MAO... and Cuifang KUANG|Show fewer author(s)
In order to achieve fast and high-precision measurement and analysis of stress birefringence in optical materials and optical components， this paper proposes a method for the two-dimensional distribution measurement of stress birefringence based on based on dual photoelastic modulators cascade difference frequency moduIn order to achieve fast and high-precision measurement and analysis of stress birefringence in optical materials and optical components， this paper proposes a method for the two-dimensional distribution measurement of stress birefringence based on based on dual photoelastic modulators cascade difference frequency modulation. Two photoelastic modulators operating at different frequencies are cascaded to form a novel polarimetry. The retardance and fast axis azimuth of the stress birefringence are loaded into the modulation signals. Employing digital phase-locked technology， the fundamental and differential frequency harmonic terms are extracted， and then the two parameters of stress birefringence are solved out. According to the principle analysis， the system is developed， and the initial offset value of the system are calibrated. A wave plate is used to measure accuracy and repeatability， and the stress birefringence distribution measurement experiment is completed with a BK7 glass sample. The experimental results show that the repeatability of fast axis azimuth and retardance is 0.01° and 0.02 nm， respectively， and the measurement time of a single point data is less than 200 ms. The scheme realizes high-speed， high-precision and high-repeatability stress birefringence measurement. The method demonstrate the ability to measure the two-dimensional distribution of stress birefringence. This provides an effective means for the analysis and evaluation of birefringence measurement of optical materials such as wave plates， glass or crystals..
Optics and Precision Engineering
- Publication Date: Sep. 25, 2023
- Vol. 31, Issue 18, 2647 (2023)
© Copyright 2018-2021 | Chinese Laser Press. All Rights Reserved 沪ICP备15018463号-20