Contour extraction is a fundamental problem in computervision, and widely applied to visual projects. However, for some scenes with complex background or objects to be extracted are similar to background, traditional...
详细信息
ISBN:
(纸本)9781538695944
Contour extraction is a fundamental problem in computervision, and widely applied to visual projects. However, for some scenes with complex background or objects to be extracted are similar to background, traditional methods seem powerless. In this paper, we present a new approach to tackle this problem. Given an image sequence, the approach is decomposed into two main phases: background modeling, contour extracting. In the first phase, Gaussian mixture distribution is applied to build the background model in dynamic environment. In the second phase, we map the segmented foreground pixels to a two-dimensional point set, through which we construct a geometric graph based on Euclidean distance, and then all border nodes are connected to generate the contour using the orientation information. Actual project results show that the approach is capable of generating contours with high accuracy and strong robustness.
imageprocessing is generally presented in a signal and systems environment which is disregarded to some business oriented Informatics Engineering students as heavy mathematical and boring. Only a few of these student...
详细信息
As an important part of connecting engine, air filter, engine, cooling system and automobile air-conditioning system, automotive hose is widely used in automobile. Therefore, the determination of the surface quality o...
详细信息
As an important part of connecting engine, air filter, engine, cooling system and automobile air-conditioning system, automotive hose is widely used in automobile. Therefore, the determination of the surface quality of the hose is particularly important. This research is based on machine vision technology, using HALCON algorithm for the processing of the hose image, and identifying the surface defects of the hose. In order to improve the detection accuracy of visual system, this paper proposes a method to classify the defects to reduce misjudegment. The experimental results show that the method can detect surface defects accurately.
High-resolution surface topography measurement using two line scan CCDs was proposed for full-field optical inspection of surface geometry and defects. Based on traditional stereo vision technique, a new configuration...
详细信息
High-resolution surface topography measurement using two line scan CCDs was proposed for full-field optical inspection of surface geometry and defects. Based on traditional stereo vision technique, a new configuration of dual linear CCD system was constructed and a novel algorithm for correspondence matching based on digital image correlation (DIC) technique was proposed for the determination of 3D coordinates. Depending on the desired resolution, a user-defined template was employed in the algorithm to accurately select the correspondence pairs in the line scan image. The OpenGL (TM) database was used to reconstruct the surface topography from the corresponding matched points. Experimental results show that the accuracy and repeatability in the sub-micrometer scale were established with the proposed technique, which can be further improved depending on the pixel resolution of the line CCD. Since larger specimen can be inspected at high-speed by the use of larger arrays with smaller pixel size, the dual vision system can be easily used for high-resolution 3D profile measurement and defect inspection in-line.
How to ensure the safety of a locomotive is a crucial problem, when it moves in a curved railway at night. For solving this problem, a new adaptive locomotive headlamp system based on monocular vision is proposed in t...
详细信息
ISBN:
(纸本)9789811079863;9789811079856
How to ensure the safety of a locomotive is a crucial problem, when it moves in a curved railway at night. For solving this problem, a new adaptive locomotive headlamp system based on monocular vision is proposed in this paper. The method consists of two key steps. First, a single-camera is provided to capture the tracks in front of the locomotive, and an imageprocessing algorithm is proposed to obtain the key related parameters. Next, combined with the position feedback signal of the headlamps and the parameters obtained in the previous step, an adaptive control method is proposed to rotate the headlamps. By this way, the light emitted from the headlamps will always be on the axis of the railway, thus ensuring the safety of the locomotive.
Road traffic and traffic congestion are major problems worldwide. To avoid such problems surveillance is the most economical technique for monitoring road traffic. Traffic monitoring has become vital to make sure quic...
详细信息
Traffic management is one of the challenging issues that need to be addressed by any urban area, one of which is Tagbilaran City, located at the Province of Bohol, Philippines. Thus, the need to employ measures such a...
详细信息
ISBN:
(纸本)9781450377201
Traffic management is one of the challenging issues that need to be addressed by any urban area, one of which is Tagbilaran City, located at the Province of Bohol, Philippines. Thus, the need to employ measures such as traffic surveillance systems is imperative. In such kind of system, vehicle detection is a basic functionality and in this paper, a YOLO-based model is developed to detect a tricycle, which is a unique kind of public transportation. Training of the model is done on images of tricycles, extracted from actual traffic videos of selected intersections and the performance of the model is measured using the average precision. In this case, a 37.91% average precision is generated for tricycles. Increasing the number of annotated images of tricycles in the training dataset will produce a more precise detection model and including more of the other types of common vehicles will generate data that will support any traffic reduction measure.
An examination is made of the robustness of correspondence-based approaches to structure from motion. It is proven in an algorithm-independent way, that small absolute errors in image displacements cause absolute erro...
详细信息
ISBN:
(纸本)0818620579
An examination is made of the robustness of correspondence-based approaches to structure from motion. It is proven in an algorithm-independent way, that small absolute errors in image displacements cause absolute errors in rotational motion parameters significant enough to lead to large relative errors in the determination of environmental depth. Even if the exact motion parameters are known a priori small errors in image displacements still lead to large errors in depth for environmental points whose distance from the camera is greater than a few multiples of the total translation in depth of the camera. In order to clarify the many robustness issues that are raised, a depth determination algorithm is developed and applied to dynamic image sequences of natural outdoor scenes.
In order to effectively extract the characteristic information of microscopic image feature to Chinese herbal medicines (CHM), and improve the recognition accuracy automatically, a novel algorithm using Pulse Coupled ...
详细信息
ISBN:
(纸本)9781538649916
In order to effectively extract the characteristic information of microscopic image feature to Chinese herbal medicines (CHM), and improve the recognition accuracy automatically, a novel algorithm using Pulse Coupled Neural Networks (PCNN) is put forward. Firstly, the PCNN model is introduced from suitable for processingimage of biological tissue. Secondly, the characteristic of time series with PCNN imageprocessing is formed, and transformed into the feature of one dimensional entropy series, which can behalf the image inherent characteristics. Finally, the automatic identification is taken to the extracted image entropy sequence feature. The experimental results show that the entropy sequence feature has the ability of anti-geometric distortions, the novel method have characteristics of simple extraction approach, little extraction parameter, easy implementation, higher accurate recognition ratio and strong robustness.
A model is described for image segmentation that tries to capture the low-level depth reconstruction exhibited in early human vision, giving an important role to edge terminations. The problem is to find a decompositi...
详细信息
ISBN:
(纸本)0818620579
A model is described for image segmentation that tries to capture the low-level depth reconstruction exhibited in early human vision, giving an important role to edge terminations. The problem is to find a decomposition of the domain D of an image that has a minimum of disrupted edges--junctions of edges, crack tips, corners, and cusps--by creating suitable continuations for the disrupted edges behind occluding regions. The result is a decomposition of D into overlapping regions R124...24Rn ordered by occlusion, which is called the 2.1-D sketch. Expressed as a minimization problem, the model gives rise to a family of optimal contours, called nonlinear splines, that minimize length and the square of curvature. These are essential in the construction of the 2.1-D sketch of an image, as the continuations of disrupted edges. An algorithm is described that constructs the 2.1-D sketch of an image, and gives results for several example images. The algorithm yields the same interpretations of optical illusions as the human visual system.
暂无评论