- 标题
- 摘要
- 关键词
- 实验方案
- 产品
-
Online weld pool contour extraction and seam width prediction based on mixing spectral vision
摘要： In this paper, based on gas–metal–arc welding (GMAW), we used a passive vision sensing system and proposed a double-path imaging method to capture weld pool images, and empirically and theoretically demonstrated the optimal bands. According to the mixed spectra of self-emitted radiation of the weld pool and the arc spectra, we selected 660-nm narrowband and 850-nm long-pass as the system’s working bands. Two cameras with 660-nm narrowband filter and 850-nm long-pass filter were used to capture weld pool images at the background level through a synchronous acquisition equation and weld pool images with high signal-to-noise ratio were obtained. After image registration, we used Gradient and Gray-based Neighbor Superpixel Merging (GNSM) method to extract the contour of weld pool image. Comparing with other algorithms, the proposed algorithm can obtain an effective and accurate contour of the weld pool image. Then we proposed an online seam width prediction method before seam formation which is based on the contour of the weld pool image. We used Gaussian distribution to fit the pixel width of the contour and the corresponding seam width measured by three-dimensional reconstruction. By comparatively analyzing the fitting deviation and the actual measurement results, we concluded that the deviation of weld seam width prediction was within 0.20 mm.
关键词： Three-dimensional reconstruction,Superpixel segmentation,Seam width prediction,Vision sensing system,Double-band imaging,Contour of weld pool
更新于2025-09-04 15:30:14
-
[IEEE 2017 IEEE International Conference on Computational Intelligence and Computing Research (ICCIC) - Coimbatore (2017.12.14-2017.12.16)] 2017 IEEE International Conference on Computational Intelligence and Computing Research (ICCIC) - An Automated Tool to Segment Blood Vessel from RGB Retinal Images
摘要： This paper proposes an approach to examine the condition of retinal blood vessel in the benchmark retinal images. An automated tool to segment the blood vessel is constructed with multi-scale matched filter along with the bat algorithm (BA). The chief motivation behind the use of BA is to discover optimal filter parameters to attain better segmentation accuracy. The proposed tool is tested using the benchmark DRIVE dataset. The experimental work is implemented using the Matlab software and the superiority of the constructed tool is appraised by computing the well known picture similarity and statistical measures. Finally, the superiority of the proposed approach is justified with a relative analysis against similar procedures available in the literature. The results of this paper confirm that, the implemented tool is efficient in extracting the blood vessel contrast to other techniques existing in the literature.
关键词： Retinal image,segmentation,blood vessel,bat algorithm,similarity measure
更新于2025-09-04 15:30:14
-
[IEEE 2018 IEEE 14th International Conference on Intelligent Computer Communication and Processing (ICCP) - Cluj-Napoca (2018.9.6-2018.9.8)] 2018 IEEE 14th International Conference on Intelligent Computer Communication and Processing (ICCP) - A Deep Learning Approach For Pedestrian Segmentation In Infrared Images
摘要： Semantic segmentation in the context of traffic scenes has been vastly explored using different architectures for deep convolutional networks and color images. In the case of infrared images there is place for improvement and scientific contributions mainly due to the lack of data sets that contain baseline segmentations in the infrared domain. This paper proposes a method for real time infrared pedestrian segmentation using ERFNet. Within the context of the proposed method we study the effect of different basic image enhancement techniques on the performance of the segmentation. We enhance an existing dataset of infrared images with ground truth segmentations for pedestrians. Our experiments show that the proposed method is accurate and appropriate for real time applications.
关键词： pedestrian segmentation,ERFNet,infrared images,deep learning,image enhancement
更新于2025-09-04 15:30:14
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - The Effect of Focal Loss in Semantic Segmentation of High Resolution Aerial Image
摘要： The semantic segmentation of High Resolution Remote Sensing (HRRS) images is the fundamental research area of the earth observation. Convolutional Neural Network (CNN), which has achieved superior performance in computer vision task, is also useful for semantic segmentation of HRRS images. In this work, focal loss is used instead of cross-entropy loss in training of CNN to handle the imbalance in training data. To evaluate the effect of focal loss, we train SegNet and FCN with focal loss and confirm improvement in accuracy in ISPRS 2D Semantic Labeling Contest dataset, especially when (cid:13) is 0.5 in SegNet.
关键词： deep learning,focal loss,semantic segmentation,CNN
更新于2025-09-04 15:30:14
-
[IEEE 2018 2nd International Symposium on Multidisciplinary Studies and Innovative Technologies (ISMSIT) - Ankara, Turkey (2018.10.19-2018.10.21)] 2018 2nd International Symposium on Multidisciplinary Studies and Innovative Technologies (ISMSIT) - An Efficient Retinal Blood Vessel Segmentation using Morphological Operations
摘要： The structure of retinal vessel carries information about many diseases. It is difficult to analyze this complex structure by human eye. Additionally, it has time-consuming process. In this study, an extremely lower complex and more successful retinal blood vessel segmentation method is proposed via using morphological operators. Colorful retinal images are divided into red, green and blue channels. Green channel is preferred for segmentation on the account of including clear details about retinal vessels. Then, adaptive threshold with 5x5 Gaussian window is applied in order to obtain clean vessel geometry. In the next step, retinal image is sharpened and then, 3x3 wiener filter is applied to it. After wiener filter, some noise in the image decreases but retinal image pixels soften. Therefore, Otsu thresholding is applied to softened images. Finally, morphological operation is performed on gray level images. The proposed method is implemented on test images in DRIVE database. The process time of our method is 0.7-0.8 second and it is faster than other methods. 95,61% accuracy, 85.096% sensitivity and 96.33% specificity rates are obtained.
关键词： image texture analysis,Biomedical image processing,image denoising,segmentation,image edge detection
更新于2025-09-04 15:30:14
-
[IEEE 2018 IEEE 7th Global Conference on Consumer Electronics (GCCE) - Nara, Japan (2018.10.9-2018.10.12)] 2018 IEEE 7th Global Conference on Consumer Electronics (GCCE) - Iterative Superpixel Segmentation Based on Color Differences
摘要： In order to find the undiscovered delicate seams, we propose a simple method that iteratively finds the seams from coarse to fine. We utilize the edge energy map to constraint the seam boundary, which is simple and effective to implement by the Sobel filters. Then, we iteratively find the optimal horizontal and vertical seams. The optimal seams are found by using the dynamic programming technique. The proposed method also checks the color homogeneity and the minimum width of superpixel in the iteration step, so it can find undiscovered seams. This algorithm is tested on Berkeley Segmentation Dataset (BSD300), and it is competitive in accuracy with other algorithms.
关键词： superpixel,dynamic programming,segmentation
更新于2025-09-04 15:30:14
-
[IEEE 2018 IEEE International Conference on Intelligent Transportation Systems (ITSC) - Maui, HI, USA (2018.11.4-2018.11.7)] 2018 21st International Conference on Intelligent Transportation Systems (ITSC) - Real-time Stereo Reconstruction Failure Detection and Correction using Deep Learning
摘要： This paper introduces a stereo reconstruction method that besides producing accurate results in real-time, is capable to detect and conceal possible failures caused by one of the cameras. A classification of stereo camera sensor faults is initially introduced, the most common types of defects being highlighted. We next present a stereo camera failure detection method in which various additional checks are being introduced, with respect to the aforementioned error classification. Furthermore, we propose a novel error correction method based on CNNs (convolutional neural networks) that is capable of generating reliable disparity maps by using prior information provided by semantic segmentation in conjunction with the last available disparity. We highlight the efficiency of our approach by evaluating its performance in various driving scenarios and show that it produces accurate disparities on images from Kitti stereo and raw datasets while running in real-time on a regular GPU.
关键词： error correction,convolutional neural networks,stereo reconstruction,failure detection,semantic segmentation
更新于2025-09-04 15:30:14
-
ChipNet: Real-Time LiDAR Processing for Drivable Region Segmentation on an FPGA
摘要： This paper presents a field-programmable gate array (FPGA) design of a segmentation algorithm based on convolutional neural network (CNN) that can process light detection and ranging (LiDAR) data in real-time. For autonomous vehicles, drivable region segmentation is an essential step that sets up the static constraints for planning tasks. Traditional drivable region segmentation algorithms are mostly developed on camera data, so their performance is susceptible to the light conditions and the qualities of road markings. LiDAR sensors can obtain the 3D geometry information of the vehicle surroundings with high precision. However, it is a computational challenge to process a large amount of LiDAR data in real-time. In this paper, a CNN model is proposed and trained to perform semantic segmentation using data from the LiDAR sensor. An efficient hardware architecture is proposed and implemented on an FPGA that can process each LiDAR scan in 17.59 ms, which is much faster than the previous works. Evaluated using Ford and KITTI road detection benchmarks, the proposed solution achieves both high accuracy in performance and real-time processing in speed.
关键词： Autonomous vehicle,LiDAR,FPGA,road segmentation,CNN
更新于2025-09-04 15:30:14
-
[IEEE 2018 25th IEEE International Conference on Image Processing (ICIP) - Athens, Greece (2018.10.7-2018.10.10)] 2018 25th IEEE International Conference on Image Processing (ICIP) - Improving Time-of-Flight Sensor for Specular Surfaces with Shape from Polarization
摘要： Time-of-Flight (ToF) sensors can obtain depth values for diffuse objects. However, the essential problem is that these sensors cannot receive active light from specular surfaces due to specular reflections. In this paper, we propose a new depth reconstruction framework for specular objects that combines ToF cues and Shape from Polarization (SfP). To overcome the ill-posedness of SfP with a single view, we integrate superpixel segmentation with planarity constraints for every superpixel. Experimental results demonstrate the effectiveness of the depth reconstruction algorithm for both controlled environment data and real vehicle data in a parking area.
关键词： Shape from Polarization,specular surface,Time-of-Flight sensor,single view reconstruction,superpixel segmentation
更新于2025-09-04 15:30:14
-
Novel Segmentation of Iris Images for Biometric Authentication Using Multi Feature Volumetric Measure
摘要： The aim of the research is to improve the efficiency of biometric authentication using different features of iris image. The biometric authentication and verification has become more popular where the authentication is more essential in many organizations. There are many approaches has been discussed to segment the iris image and to perform verification but suffers with the problem of accuracy in feature extraction and segmentation. To resolve such problems and to improve the efficiency of iris segmentation and recognition, we propose a novel segmentation algorithm which uses multi level filter which removes the eyelids and eyelash features and performs the edge detection to identify the inner and outer eye regions. Once the regions has been identified then, we compute various measures like the size of inner and outer eyes and extract the features of both and convert them in to feature vectors. The generated feature vectors are used to perform classification in biometric authentication approach. The multi feature volumetric measure is computed on the feature vector of each eye image where the feature vector has various features like the size of both inner and outer eyes, width and height, the original binary features, the number of binary ones and the number of pixels damaged by any form of disease and so on. Based on these features the MFVM is computed to classify the iris image towards a big data set of biometric features to perform authentication. The proposed method has improved the efficiency of iris segmentation and improved the efficiency of iris recognition based biometric authentication. Also the approach has reduced the time complexity and improved the efficiency also.
关键词： MFVM,iris segmentation,iris recognition,Biometric authentication,multi level filtering
更新于2025-09-04 15:30:14