The number of concrete structures that are more than 50 years old is increasing. Since these concrete structures must be inspected periodically, automatic crack inspection methods are utilized, such as image processin...
详细信息
ISBN:
(纸本)9784907764678
The number of concrete structures that are more than 50 years old is increasing. Since these concrete structures must be inspected periodically, automatic crack inspection methods are utilized, such as imageprocessing to find shapes of cracks on the concrete surface. The crack inspection result needs to be corrected and confirmed off-line by an operator to verify that there are no misseddefects or false positives. In this paper, we developed a crack inspection support system using a head-mounteddisplay (HMd). The proposed system consists of building 3d models by spatial sensing, detecting and matching features of cracks from capturedimages, and finally superimposing and projecting detected cracks onto the real structure surface in a mixed reality (MR) space. Through experiments, we confirmed that the proposed system could acquire 3ddata of concrete structures that can be measured easily in real time and onsite. In addition, it is possible for the operator to confirm the crack measurement analysis by comparing measured results with the actual structure in the MR environment. Thus, the proposed system unifies the previously separated processes of 3ddata creation and analysis used in concrete structure inspection.
The following topics are dealt with: learning (artificial intelligence); medical imageprocessing; convolutional neural nets; feature extraction; neural nets; image segmentation; electroencephalography; image classifi...
详细信息
The following topics are dealt with: learning (artificial intelligence); medical imageprocessing; convolutional neural nets; feature extraction; neural nets; image segmentation; electroencephalography; image classification; medical signal processing; 5G mobile communication.
Geological section measurement is a necessary task for regional geological survey. Traditional method using tapeline together with compass has many limitations such as multiple manual measurements along different orie...
详细信息
ISBN:
(纸本)9781728123264
Geological section measurement is a necessary task for regional geological survey. Traditional method using tapeline together with compass has many limitations such as multiple manual measurements along different orientation across different structures, visual conditions, tapeline bending, manual measurement error and second projection defects. Point coordinates method recently established have allowed us to use points with X, Y, Z values to obtain terrain control profiles. However, it is difficult to adequately reveal the spatial and tectonic characteristics of outcrops in arbitrary orientation using sparse GPS coordinates. In order to obtain more realistic geological structural features, this paper propose a new section measurement method that utilizes high-density point clouddata with X, Y, Z values obtained by LidAR and high-resolution digital images with RGB values. Based on digital imageprocessing technology, the high-resolution digital images are fused with the high-density point clouddata. And then, a digital outcrop model (dOM) is generated, which contains spatial coordinate information, RGB values of any points of outcrops and surface morphological features in a 3d space. Finally, a geological section in arbitrary orientations can be generated efficiently and quickly under the point coordinate method after the dOM ortho-projected. Because the spatial resolution of point clouddata reaches millimeter scale, the outcrop section parameters extracted by this approach, such as fault displacement, stratum thickness, are more precise than manual measurements (generally only meter scale). Furthermore, LidAR and photogrammetry are both non-contact measurement technologies, which greatly reduce the potential risk of geologists in the field environments like cliff.
depth cameras have enhanced the environment perception for robotic applications significantly. They allow to measure true distances and thus enable a 3dmeasurement of the robot surroundings. In order to enable robust...
详细信息
ISBN:
(纸本)9781450372886
depth cameras have enhanced the environment perception for robotic applications significantly. They allow to measure true distances and thus enable a 3dmeasurement of the robot surroundings. In order to enable robust robot vision, the objects recognition has to handle rotateddata because object can be viewed from different dynamic perspectives when the robot is moving. Therefore, the 3ddescriptors used of object recognition for robotic applications have to be rotation invariant and implementable on the embedded system, with limited memory and computing resources. With the popularization of the depth cameras, the Histogram of Gradients (HOG) descriptor has been extended to recognize also 3d volumetric objects (3dVHOG). Unfortunately, both version are not rotation invariant. There are different methods to achieve rotation invariance for 3dVHOG, but they increase significantly the computational cost of the overall dataprocessing. Hence, they are unfeasible to be implemented in a low cost processor for real-time operation. In this paper, we propose an object pose normalization method to achieve 3dVHOG rotation invariance while reducing the number of processing operations as much as possible. Our method is based on Principal Component Analysis (PCA) normalization. We tested our method using the Princeton Modelnet10 dataset.
Blind quality assessment is a challenging issue since the evaluation is done without access to the reference nor any information about the distortion. In this work, we propose an objective blind method for the visual ...
详细信息
Blind quality assessment is a challenging issue since the evaluation is done without access to the reference nor any information about the distortion. In this work, we propose an objective blind method for the visual quality assessment of 3d meshes. The method estimates the perceived visual quality using only information from the distorted mesh to feed pre-traineddeep convolutional neural networks. The input data is prepared by rendering 2d views from the 3d mesh and the corresponding saliency map. The views are split into small patches of fixed size that are filtered using a saliency threshold. Only the salient patches are selected as input data. After that, three pre-traineddeep convolutional neural networks are used for the feature learning: VGG, Alexnet and Resnet. Each network is fine-tuned and estimates separately a quality score. Finally, a weighted sum is applied to compute the final objective quality score. Extensive experiments are conducted on a publicly available database, and comparisons with existing methods demonstrate the effectiveness of the proposed method according to the correlation with subjective scores.
Three-dimensional (3d) images are widely used in the medical field (e.g., CT, MRI). In osteoarthritis research, 3d magnetic resonance imaging (MRI) provides a noninvasive way to study soft-tissue structures including ...
详细信息
Three-dimensional (3d) images are widely used in the medical field (e.g., CT, MRI). In osteoarthritis research, 3d magnetic resonance imaging (MRI) provides a noninvasive way to study soft-tissue structures including hyaline cartilage, meniscus, muscle, bone marrow lesion, etc. The measurement of those structures can be greatly improved by accurately locating the bone structure. U-net is a convolutional neural network developed for biological image segmentation using limited training data. The original U-net takes a single 2dimage as input and generates a binary 2dimage as output. In this paper, we modified the U-net model to identify the bone structure on 3d knee MRI, which is a sequence of multiple 2d slices. Instead of taking a single image as input, the modified U-net takes multiple adjacent slices as input. The output is still a single binary image which is the segmentation result of the center slice in the input sequence. By using 99 knee MRI cases, where each knee case includes 160 2d slices, the proposed model was trained, validated, and tested. The dice coefficient, similarity, and area error metrics rate were tallied to assess the performance and the quality of the testing sets. Without any post-processing of the images, the model achieved promising segmentation performance with the dice coefficient (dICE) 97.22% on the testing dataset. To achieve the best performance, diverse models were trained using different strategies including different numbers of input channels anddifferent input image sizes. The experiment results indicate that the incorporation of neighboring slices generated better segmentation performance than using the single slice. We also found that a larger image size (uncompressed) corresponds to better performance. In summary, our best segmentation performance was achieved using five adjacent neighbor slices (two left neighbors + two right neighbors + the center slice) with the original image size of 352 × 352 pixels.
In the weld quality inspection system based on machine vision and computer imageprocessing, according to the principle of laser triangulation, a line laser is used to project structured light on the weld surface, and...
In the weld quality inspection system based on machine vision and computer imageprocessing, according to the principle of laser triangulation, a line laser is used to project structured light on the weld surface, and the weldment moves uniformly in a straight line on the moving platform. At the same time, the CMOS camera captures the structured light image and transmits it to the industrial computer through Gigabit Ethernet. The image is denoised by the median filtering algorithm on the industrial computer, then the image is segmented by the maximum inter-class error (Otus) method, and the image is binarized according to the threshold value. Then the center of the stripe is extracted by constructing four direction templates, and the 3d model is reconstructed by curve fitting. After the completion of the imageprocessing, the geometric size of the weld is measured by mathematical formula. After the weld seam measurement is completed, the weld seam is transversely cut off at the same position, measured by manual method, and then the data measured by the two methods are compared. Through repeatedmeasurements and comparisons of weld width, residual height and back forming, the precision of the system meets the practical requirements. The weld forming varies with the welding parameters within a certain range, but changes little and has good stability.
暂无评论