Many industrial manufacturing processes are not well understood and are treated as `black art' with few experts able to control the process and ensure product quality. However, modern manufacturing companies are f...
详细信息
Many industrial manufacturing processes are not well understood and are treated as `black art' with few experts able to control the process and ensure product quality. However, modern manufacturing companies are finding it increasingly difficult to compete in the global marketplace without better process understanding and control. Automated inspection systems for general manufacturing have become more feasible through technical advances, primarily in sensor and computing technology. However, these systems have been used almost exclusively for the detection and subsequent removal of well defined, discrete defects from the product;thus guaranteeing high quality for the customer. This paper describes a larger opportunity to affect operations by employing web inspection techniques to dynamically analyze manufacturing conditions rather than just detecting the presence of defective material. One can then keep the process under better control, thereby eliminating defects, ensuring product quality, and optimizing manufacturing time on the production line. Specific image and data processing techniques will be illustrated including product uniformity metrics, automatic determination of thresholds for blob analysis, and localization of repeating defects within production data. The benefit of these techniques will be demonstrated through `real-world' examples of web-based manufactured products.
imageprocessing plays an important role in ITS. Since a traffic environment heavily depends on visual information, machinevision is a must in ITS. There are many active research and development programs applying ima...
详细信息
imageprocessing plays an important role in ITS. Since a traffic environment heavily depends on visual information, machinevision is a must in ITS. There are many active research and development programs applying image technologies for ITS. First of all, we need a good and clear image as an input to obtain better imageprocessing results. For this purpose, resolution and dynamic range of an imaging device should be improved. Infrared and microwave imaging are attractive. A depth map from laser radar or microwave radar could be treated as an input to imageprocessing. There are many applications for imageprocessing in ITS. Typical on-board applications are: lane marker detection, obstacle detection, proximity vehicle detection, traffic signal/sign recognition, road surface monitoring, driver monitoring, etc. Roadside applications are traffic surveillance (traffic flow, travel time), incident/accident detection, license number recognition (ETC, parking lot) and parking lot surveillance.
The auroral emissions observed in the high-latitude regions encircling the magnetic poles are a key element in studying plasma physical processes in the near-Earth space, the magnetosphere. The Finnish Meteorological ...
详细信息
The auroral emissions observed in the high-latitude regions encircling the magnetic poles are a key element in studying plasma physical processes in the near-Earth space, the magnetosphere. The Finnish Meteorological Institute operates five digital all-sky cameras, which routinely monitor the auroral emissions in Northern Finland, Sweden, and Svalbard; each camera records an image of the full sky at 20 second intervals. In this paper, we develop a method that allows us to examine such a large data set by classifying the images through determining the shape skeletons of the auroral forms in each auroral image. Shape skeletons are a commonly used representation of object shapes in machinevisionapplications. Once determined, shape skeletons have the advantage that they can also be used to represent noisy or unevenly distributed data. Here we apply a skeletonising algorithm to determine the skeletons of auroras in a noisy environment. The algorithm is based on a batch mode self-organising map. The results can be further improved by implementing understanding of the auroral physics to the algorithm.
作者：
M. PittoreC. BassoA. VerriINFM-DISI
Università di Genova Italy INFM-DIFI
Università di Genova Italy INFM-DISI
Massachusetts Institute of Technology (US) Center for Biological and Computation Learning Università di Genova Italy
Support vector machines (SVM) have been recently introduced as techniques for solving pattern recognition and regression estimation problems. SVM are derived within the framework of statistical learning theory and com...
详细信息
Support vector machines (SVM) have been recently introduced as techniques for solving pattern recognition and regression estimation problems. SVM are derived within the framework of statistical learning theory and combine a solid theoretical foundation with very good performances in several applications. In this paper we describe a system able to detect, represent, and recognize visual dynamic events from an image sequence. While the events are initially detected by means of low-level visual processing, both the representation and recognition stages are performed with SVM. Therefore, the system is trained, instead of programmed, to perform the required tasks. The very good results obtained on real image sequences indicate that SVM can be profitably used for the construction of flexible and effective systems based on computer vision.
A new integrated video sensor design for applications in traffic management and control to solve problems associated with affordability and reliability was discussed. This integrated video sensor has many features suc...
详细信息
ISBN:
(纸本)9608052033
A new integrated video sensor design for applications in traffic management and control to solve problems associated with affordability and reliability was discussed. This integrated video sensor has many features such as integration of the transducers and the MVP into one compact unit improving detection reliability and lowering susceptibility to EMI. The sensor also lowered installation costs, reduced installation time and enabled rapid deployment making the system more readily portable. The new communication architecture offered optimal routing of the machinevision detection results, full motion video as well as digital imagery and supervisory control.
A multi-configuration stereoscopic vision-system is described. We present the mechanical and electronic features of the stereoscopic head together with the hardware. We describe the methods of imageprocessing impleme...
详细信息
A multi-configuration stereoscopic vision-system is described. We present the mechanical and electronic features of the stereoscopic head together with the hardware. We describe the methods of imageprocessing implemented for overcoming the optical quality drawback of the low-cost CCD mini-cameras. Algorithms and results for the high-level image representation and 3D reconstruction are presented. The structural characteristics, the low-cost and the precision of the stereo system built, allow its utilisation in domestic mobile robotic applications.
Most vision systems require the use of imageprocessingapplications that are highly dependent on the efficiency of edge detection techniques. These techniques are commonly implemented by applying an edge enhancement ...
详细信息
Most vision systems require the use of imageprocessingapplications that are highly dependent on the efficiency of edge detection techniques. These techniques are commonly implemented by applying an edge enhancement method followed by a thresholding point process. Most of these techniques are based on convolution algorithms that have a time complexity of O(n/sup 2/) when the picture has size n/spl times/n. In order to reduce this time complexity, an improved solution depends on the reduction of the problem space. Such a reduction was recently achieved by a new method, named GALE, which combines the random search mechanisms of genetic algorithms with linear time methods. In this paper, a refinement of the GALE method is accomplished by introducing an iterative process that selectively eliminates from the population of the genetic algorithm those pixels that were previously identified as part of an edge. Experimental results show the improved performance of this method.
Many imaging systems with industrial and machinevisionapplications now require sensors with a combination of >1 Mpixel revolution and >30 frames/s rate. This must be achieved at low cost and with minimum impac...
详细信息
Many imaging systems with industrial and machinevisionapplications now require sensors with a combination of >1 Mpixel revolution and >30 frames/s rate. This must be achieved at low cost and with minimum impact on system complexity or power consumption. Reported CCDs have this level of performance. However, the inherent analog nature of CCDs leads to a requirement for driving large external loads at high frequency from the image sensor. CMOS active pixel sensors (APS) are ideal for these applications due to the possibility for on-chip signal processing (especially analog-to digital converters [ADC]) integration and inherent low power consumption. This advantage is especially important in large-format high speed imaging applications. Other CMOS image sensors reported to date have either targeted small-format applications or are low frame rate devices. This 1280(H)x 720(V) CMOS APS capable of 60 frames/s operation includes on-chip, 10b, column-parallel, self-calibrating ADCs with 9b performance. Four 10b wide digital output ports support a data rate of 55.3 Mpixels/s. The sensor is in a standard 0.5/spl mu/m 2P2M technology and has power consumption <250 mW.
Describes the enhancement of Xvision, a large library of C++ code for real-time visionprocessing, into Fvision (pronounced "fission"), a fully-featured domain-specific language (DSL) embedded in Haskell. Th...
详细信息
Describes the enhancement of Xvision, a large library of C++ code for real-time visionprocessing, into Fvision (pronounced "fission"), a fully-featured domain-specific language (DSL) embedded in Haskell. The resulting prototype system substantiates the claims of increased modularity, effective code reuse and rapid prototyping that characterize the DSL approach to systems design. It also illustrates the need for judicious interface design: relegating computationally expensive tasks to Xvision (pre-existing C++ components) and leaving modular compositional tasks to Fvision (Haskell). At the same time, our experience demonstrates how Haskell's advanced language features (specifically, parametric polymorphism, lazy evaluation, higher-order functions and automatic storage reclamation) permit a rapid DSL design that is itself highly modular and easily modified. Overall, the resulting hybrid system exceeded our expectations: visual tracking programs continue to spend most of their time executing low-level imageprocessing code, while Haskell's advanced features allow us to quickly develop and test small prototype systems within a matter of a few days, and to develop realistic applications within a few weeks.
The authors have applied a systems philosophy to the computer vision problem, and they have designed a system called O-NAV (Object NAVigation) that can harness all of the computer vision technology to date, and combin...
详细信息
The authors have applied a systems philosophy to the computer vision problem, and they have designed a system called O-NAV (Object NAVigation) that can harness all of the computer vision technology to date, and combine these approaches into one integrated system. In O-NAV, no one sub-component bears the burden of the problem. In other words, it is not expected that algorithms alone will solve the computer vision problem. If we choose effective sensing that inherently performs some level of scene discrimination at the phenomenology level, algorithms will be handed a partially analyzed scene before they ever encounter the raw image data. The algorithms have been designed to exploit an optimized processing hardware infrastructure, to maximize computation for the "real-time" application of autonomous robot navigation.
暂无评论