To solve the conflicting problem of high natural frequency and low sensitivity of geophones in the exploration and detection of medium-high frequency seismic signals, a fiber grating medium-high frequency seismic geophone based on a lever amplification mechanism is proposed. One end of the inertial mass block is connected to the lever amplification mechanism, and the fiber grating is vertically fixed on the amplification end of the lever and the base using a two-point packaging method. The lever amplification mechanism can produce mechanical amplification at the output end of the detector, increase the axial strain of the fiber grating, and increase the response to weak medium-high frequency signals in the formation, thereby improving the sensitivity of the detector. At the same time, the mechanical structure of the detector uses integrated processing technology, which can improve the rigidity and stability of the mechanical structure. Experimental tests show that the natural frequency of the detector is 526 Hz, the sensitivity is 119.1 084 pm/g in the frequency range of 20—260 Hz, and it has good lateral anti-interference performance. It provides certain technical ideas and directions to study the medium-high frequency seismic geophone.

Addressing the challenges associated with text detection in complex natural scenes, this paper presents a novel scene text detection method that employs a dual-attention and multi-scale feature fusion strategy. By introducing the dual-attention fusion mechanism, the correlation between text feature channels is strengthened, leading to an overall improvement in detection performance. Furthermore, considering the potential loss of semantic information resulting from up-and-down sampling of deep feature maps, a hollow convolutional multi-scale feature fusion pyramid is introduced. This approach adopts a dual fusion mechanism to enhance semantic features and overcome the impact of scale variations. To address the issues of semantic conflict and limited representation of multi-scale features resulting from the fusion of information with different densities, an innovative multi-scale feature fusion module (MFFM) is introduced. In addition, the feature refinement module (FRM) is introduced for the problem of small text that is easily masked by conflicting information. The experiments show the effectiveness of our method for text detection in complex scenes with F-values of 85.6%, 87.1% and 86.3% on three datasets, CTW1500, ICDAR2015, and Total-Text.

Aiming at the problems of inaccurate feature extraction and insufficient use of effective features in existing dynamic scene image deblurring methods, this paper proposes a dynamic scene image deblurring network based on two-branch feature extraction and cyclic refinement. The whole network consists of feature extraction network, cyclic refinement network (CRN) and image reconstruction (IR). Among them, the feature extraction network includes the extraction of detail and contour features (CFs) of the blurred image, using the residual unit as the basic unit of the feature extraction network. The cyclic refinement network refines the feature map by alternately fusing contour features and detail features (DFs) to obtain the refinement features (RFs) of the blurred image. Finally, in the image reconstruction stage, the contour and detail features are reused and combined with the residual learning strategy to fuse the contour features, detail features and refined features step by step, and then the clear image is reconstructed by nonlinear mapping. The experimental results on the widely used dynamic scene blurring dataset GOPRO show that the average peak signal to noise ratio (PSNR) of this method reaches 31.86, and the average structure similarity (SSIM) reaches 0.947 3. The images restored by the proposed method contain rich details and have better deblurring effect. The proposed method is superior to the comparison method in terms of objective evaluation index and subjective visual effect.

Currently grassland environment is complex, pastures are scattered and have little difference in color from the background. It isn't achieved the efficient and accurate segmentation. Therefore, this paper proposes a novel lightweight and multi-scale DeeplabV3+ network (LMS-DeeplabV3+). The network uses DeeplabV3+ as the base network, and first selects the lightweight MobilenetV2 as the backbone network for initial feature extraction and adjust its configuration to suit the pasture segmentation task; secondly the depth separable convolution is used instead of normal convolution in both the enhanced feature extraction and decoding modules to lighten the network; in addition, the dense atrous spatial pyramid pooling (DASPP) module is used to capture a larger sensory field and enhance the interaction among features; the convolutional block attention module (CBAM) is also introduced to reassign weights to enhance feature extraction. Experiments show that the proposed new network improves mean intersection over union (mIOU) by 8.06 percentage points and mean pixel accuracy (mPA) by 6.75 percentage points compared with the original network, reduces both the computation and the number of parameters of network by more than 90%, improves the segmentation prediction speed, and performes better in all aspects compared with other mainstream segmentation networks.

Segmentation of the left ventricle(LV) using the distance regularized level set evolution(DRLSE) model causes it to be jagged and poorly segmented. To solve these problems currently faced by LV segmentation, this paper firstly uses a convolutional neural network (CNN)-based myocardial center-line detection algorithm to replace the manual initialization process of the level set method, and secondly proposes a non-zero level set-based preserving convexity LV segmentation method. Comparing the mean degree centrality of the DRLSE (level set method), deep learning method and the new method, it is found that the DC (dice coefficient) of the new method at the end-systole (ES) is 0.93, which is higher than the other methods. In addition, the mean Hausdorff distance (HD) of the new method at the end-diastolic (ED) and ES phases are 2.51 and 2.54, respectively, which is significantly smaller than those of the deep learning method and the level set method. The experimental results show that the new method can effectively improve the segmentation accuracy.

To solve the problem of feature extraction and target understanding in underwater image due to blurry underwater image and low contrast between target and background during underwater target detection, an underwater target detection algorithm based on frequency domain attention is presented. Firstly, the method transforms the training set image into the frequency domain, and uses the low frequency feature guiding suite (LFGS) to calculate the frequency component. Then the component will be applied as a parameter to the low frequency feature extraction model (LFM) to better extract the low frequency features of the image. The features that fuse the low frequency information of the image are further extracted to generate the high-level features. Finally, the high-level features are input into the detection head for detection. The average accuracy is 83.35% on URPC2021 dataset, which verifies the validity of this method.

Diabetic retinopathy (DR) is currently one of the leading blinding diseases in humans. Aiming at the problems of small differences between samples and uneven class distribution in DR datasets, which restrict the improvement of grading performance, this paper proposes a classification algorithm for the fusion of attention linear features diversification (FALFD). Firstly, the improved Res2Net residual network is used as the model backbone to increase the receptive field, and further improve the ability of the network to capture feature information. Secondly, the adaptive feature diversification module (AFDM) is introduced to identify the tiny pathological features that can be resolved in the fundus images, and local features with high semantic information are obtained, which avoids the limitation of a single feature region and improves the classification accuracy. Then, the bilinear attention fusion module (BAFM) is used to increase the proportion of network weights that can identify regional features. Finally, the regularized focal loss (FL) is used to further improve the classification performance of the algorithm. On the IDRID dataset, the sensitivity and specificity are 94.20% and 97.05%, and the quadratic weighting coefficient is 87.83%, respectively. On the APTOS 2019 dataset, the quadratic weighting coefficient and the area under the receiver operating curve are 88.06% and 93.90%, respectively. The experimental results show that the algorithm has some value in the field of DR classification.

Considering that Si-based materials are easily mixed with Fe impurities and Cu impurities during the fabrication process, we establish Si models with two different impurities according to the first-principles and photoelectric response theory. The effects of impurity atoms at different interstitial sites on the energy band structure and response characteristics of Si are further compared. Results show that the mixing of the two impurities can lead to changes in the energy band structure of the silicon material, resulting in an out-of-band response and a decrease in the saturation threshold of the photosensitive unit. Specifically, when the Fe impurity occupies the tetrahedral interstitial site, the energy band structure of silicon is significantly affected, and its band gap is reduced to 0.013 eV, resulting in an out-of-band absorption peak at about 1 560 nm. The Cu impurity has an obvious effect on silicon material at the hexagonal inter-stitial site, so that the band gap disappears, and an out-of-band absorption peak appears at about 1 700 nm. In these two cases, the saturation threshold of the silicon- based photosensitive unit also decreases most significantly. When irradiated by a 1 550 nm laser, the saturation thresholds are 0.001 65 W·cm-2 and 0.002 54 W·cm-2, respectively. The analysis results provide reference for the application and development of optoelectronic devices.

Organic light-emitting materials may have an impact on the human body or the environment. In addition, they also have the disadvantages of high production cost, easy quenching of lighting, and difficult control of lighting intensity and color. Therefore, it is of great significance to explore rare earth luminescent materials with superior properties and find their application value. The rare earth ion Er3+, Eu3+ co-doped LaOF phosphors were prepared by the hydrothermal-assisted solid-phase method. The phase composition, particle size, morphology and fluorescence spectrum of LaOF∶Eu3+ phosphors doped with Er3+ at different concentrations were characterized and analyzed by X-ray diffractometer (XRD), scanning electron microscope (SEM) and fluorescence analyzer. The results show that the precursor LaF3 of phosphors is well transformed into tetragonal LaOF after calcining at 900 ℃, and with the increase of Er3+ doped concentration, the luminescent properties of the phosphors show a good rule at the excitation wavelengths of 365 nm and 393 nm. Among them, the excitation at 365 nm shows the transition from orange light to yellow light, while the excitation at 393 nm shows the transition from orange light to magenta light. Finally, Er3+, Eu3+ co-doped LaOF phosphor is made into ink and printed into patterns, and when it is excited at 365 nm and 395 nm, the orange and purple emission is visible, respectively, which shows that they can be used in the field of anti-counterfeiting.

Aiming at the problem that low-altitude micro-UAVs pose a threat to public safety, this paper proposes a lightweight target detection model YOLOv5~~SS suitable for mobile terminals based on the you only look once v5 (YOLOv5) network. In this model, the lightweight network ShuffleNetv2 replaces the original backbone network of YOLOv5, introduces squeeze-and-excitation networks (SENet) attention mechanism, and uses soft non-maximum suppression (Soft-NMS) algorithm to improve the detection effect of dense overlapping targets. The experimental results show that the mean average precision@0.5 (mAP50) of the model for the detection of low-altitude micro-UAV on the dataset is 92.75%, the accuracy is 90.49%, and the number of parameters is 0.237 4 M. The number of floating-point operations is 0.9GFLOPS (giga floating-point operations).

The onset of subclinical keratoconus (subkc) is hidden, and existing medical equipment has limitations in diagnosis. Therefore, it is necessary to propose a detection method for diagnosing subclinical keratoconus. Studies have found that the mechanical properties of keratoconus (kc) change earlier than morphology, so screening subclinical keratoconus from the perspective of corneal biomechanics is more in line with clinical practice. This article utilizes corneal biomechanical features and uses point cloud data as network input data. Self organizing network (SO- Net) and self attention mechanism (SA) are combined to construct SOANet, which classifies keratoconus, subclinical keratoconus, and normal corneas. Firstly, a corneal visualization Scheimpflug technology (Corvis ST) was used to capture a corneal deformation video, which was processed to obtain a point cloud dataset. The point cloud data was then enhanced to achieve a balanced distribution of the three types of corneal data. Then the training set and test set were divided in a 3∶1 ratio, and the cornea was classified into two categories and three categories, respectively. The accuracy of the final model on the two categories and three categories test sets reached 98.3% and 91.26%, respectively, effectively identifying subclinical keratoconus and keratoconus. The experimental results indicate that constructing a subclinical keratoconus assisted diagnostic model using 3D point cloud data is feasible, and SOANet can effectively recognize subclinical keratoconus, and its classification performance is better than traditional models.

The metasurface has excellent performance as an ultra-thin micro-nano device. The use of metasurfaces to detect biomolecules has increasingly broad prospects. In this paper, the research progress of tunable metasurface detection of biomolecules is investigated, and the current research results are elaborated from the terahertz band, mid-infrared band and optical band. This paper focuses on the current methods of using applied voltage and changing ambient temperature to achieve tunable on the metasurface, summarizes and analyzes the performance of different biomolecules detected by the metasurface, and finally gives the opportunities and challenges for the future development of tunable metasurfaces in biomolecular detection.