- 标题
- 摘要
- 关键词
- 实验方案
- 产品
-
Multi-level Features Convolutional Neural Network for Multi-focus Image Fusion
摘要： Multi-focus image fusion is an important technique that aims to generate a single clean image by fusing multiple input images. In this paper, we propose a novel multi-level features convolutional neural network (MLFCNN) architecture for image fusion. In the MLFCNN model, all features learned from previous layers are passed to the subsequent layer. Inside every path between the previous layer and the subsequent layer, we add a 1x1 convolution module to reduce the redundancy. In our method, the source images first are fed to our pre-trained MLFCNN model to obtain the initial focus map. Then, the initial focus map is performed by morphological opening and closing operations and followed by a Gaussian filter to obtain the final decision map. Finally, the fused all-in-focus image is generated based on a weighted-sum strategy with the decision map. The experimental results demonstrate that the proposed method outperforms some state-of-the-art image fusion algorithms in terms of both qualitative and objective evaluations.
关键词： convolutional neural network,decision map,multi-focus image fusion,multi-level features
更新于2025-09-23 15:22:29
-
Laser multi-focus precision cutting of thick sapphire by spherical aberration rectification
摘要： Laser multi-focus separation technology is an effective way to cut thick, transparent and hard materials. In this paper, factors affecting the quality of multi-focus cutting sapphire were investigated. Optical simulations analysis was performed based on experimental results to analyze multi-focal laser beam propagation from air into sapphire. The results show that the focus interval is linearly dependent on refractive index. Spherical aberration due to refractive-index mismatch disperses the laser foci, leading to a reduction in laser power density which has a significant influence on the cutting quality. The aberration is strongly dependent on the refractivity of the material and the effective numerical aperture (NA). By redesigning the distribution of multiple foci to correct spherical aberration, sapphire with a thickness of 2 mm was cut successfully at a scanning speed of 4 mm/s. The roughness of the cutting section was ～2.5 μm , without obvious observed micro-cracks.
关键词： Optical simulation,Spherical aberration,Multi-focus,Sapphire
更新于2025-09-19 17:13:59
-
[Laser Institute of America ICALEO? 2017: 36th International Congress on Applications of Lasers & Electro-Optics - Atlanta, Georgia, USA (October 22–26, 2017)] International Congress on Applications of Lasers & Electro-Optics - Study on laser multi-focus separation technology for thick KDP crystal
摘要： Potassium dihydrogen phosphate (KDP) crystal is an important electro-optic material in various laser system, and belongs to very difficult-to-cut material, especially for thick crystals. In this study, a laser multi-focus separation technology (LMFS) of thick KDP crystal has been developed by skillfully combining femtosecond laser and LMFS optical system. In this way, the uniformity of temperature and thermal stress distributions along crystal thickness and the utilization efficiency of the laser energy are greatly improved. A penetration crack along crystal thickness could be formed and its propagation direction could be controlled to achieve safe and high-quality separation. The separating thickness of LMFS (50 mm) is at least 4 times thicker than that of existing laser separating technology, and the separating efficiency of LMFS (200 mm/s) is at least 20 times faster than that of traditional mechanical method. The generation mechanism of multi-focus was expounded by optical analysis and design, and verified by an established LMFS optical system. A thick KDP crystal with the thickness of 50 mm was separated successfully by LMFS. The quality of the separated sidewall surface was clean and flat (roughness of 10.857 nm, flatness of 3.5389 mm) without any contamination, underground damage and edge fragmentation. The experimental results proved the feasibility of LMFS, and are in good agreement with the theoretical analysis.
关键词： femtosecond laser,KDP crystal,optical system,laser multi-focus separation technology,thermal stress
更新于2025-09-16 10:30:52
-
[IEEE 2018 IEEE 3rd International Conference on Image, Vision and Computing (ICIVC) - Chongqing (2018.6.27-2018.6.29)] 2018 IEEE 3rd International Conference on Image, Vision and Computing (ICIVC) - Multi-Focus Image Fusion Using Block-Wise Color-Principal Component Analysis
摘要： In this work, multi-focus image fusion method has been proposed by using color-principal component analysis (C-PCA). Proposed method consists of different phases. In the first phase, both source images have been converted into three RGB color channels. In the next phase, for each channel, covariance’s has been calculated for both images. Special weights have been calculated to generate intermediate images. In the next phase, Convolution has been used with Gaussian blur to make image smooth. Zero-crossing based second order-derivative has been incorporated to calculate edges. In the last phase, images have been decomposed into blocks. Salient features information by using Laplacian of Gaussian and Spatial Frequency of each block have been used to get the fused image. Experimental results show that the proposed method performs well as compare to existing methods by using quality matrices.
关键词： spatial frequency,multi-focus image fusion,laplacian of Gaussian,principal component analysis
更新于2025-09-09 09:28:46
-
Impact of multi-focus beam shaping on the process stability
摘要： Unwanted process defects occur during laser deep penetration welding. The formation mechanisms of the main process defects like spatters and pores are not completely identi?ed yet. Several methods like beam oscillation or laser power modulation are known to help reducing the occurrence of defects. These methods change the temporal energy input. It can be assumed that also beam shaping in?uences the energy input into the keyhole and has the potential to in?uence process defects as well. Due to the adapted energy input keyhole instabilities, that are assumed to in?uence the formation of the defects, can be avoided. A newly developed multi-focus beam shaping optic that creates multiple beam waists in axial direction is used in this work. The additional areas with increased intensities are used to modify the energy input in the keyhole in order to explain the mechanism of spatter formation and evaluate the potential of the axial beam shaping for defect suppression during laser deep penetration welding. It is found that the number of spatters can be reduced when the middle sections of the keyhole are illuminated with a high intensity. It can be concluded that spatters can be reduced when a keyhole collapse is avoided and a high enough energy input is guaranteed in the upper keyhole sections.
关键词： Pore and spatter formation,Laser deep penetration welding,Multi-focus optic,Axial beam shaping
更新于2025-09-04 15:30:14
-
[IEEE 2018 Innovations in Intelligent Systems and Applications Conference (ASYU) - Adana, Turkey (2018.10.4-2018.10.6)] 2018 Innovations in Intelligent Systems and Applications Conference (ASYU) - A Comparative Study for Image Fusion
摘要： Image fusion that is a process of combining two or more images to generate new features or enhance the available information is a subfield of image processing. Image Fusion is divided into four main categories; multi-view fusion, multi-modal fusion, multi-temporal fusion, multi-focus fusion. The depth of field property of the optical lens causes the view to have different depth of field because it can focus on only one region at a time. Thus, it is difficult to obtain a completely focused image. Multi-focus image fusion aims to generate a fused image whose pixels are as well focused as possible by combining multiple images that contain the same scene but have different parts of focus. In this paper, a comparative study has been carried out between four multi-focus image fusion methods that are Unique Colors (UC-IF), Discrete Cosine Transform (DCT-IF), Discrete Wavelet Transform (DWT-IF) and Principal Component Analysis(PCA-IF). The performances of these methods are evaluated with four most commonly used metrics; Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), Root Mean Square Error (RMSE) and Entropy. Although Unique Color Method is simple and easy to implement, it has been found superior to the other methods.
关键词： unique color,discrete wavelet transform,multi-focus image fusion,discrete cosine transform,Image processing,principal component analysis
更新于2025-09-04 15:30:14