混合模式识别系统研究

讨论基于多种分类方法的模块组合实现的混合模式识别系统，它不同于利用多分类器输出结果表决的集成系统．提出两个系统：一个面向印刷体汉字文本识别，另一个面向自由手写体数字识别．利用多种特征和多种分类方法的组合、部分识别信息控制混淆字判别策略以及提出的动态模板库匹配后处理方法，使系统的性能与传统单一分类器系统比较，获得明显改善．实验表明：多方法多策略混合是解决复杂和增强系统鲁棒性的一条途径

用光学阵列传感器的机器人物体分类系统

本文介绍用光学阵列传感器的机器人物体分类系统。传感器直接安装在机器人的两个手指上。被抓物体的阴影通过光导纤维传到安放在“安全区”的光敏元件上。计算机识别物体的轮廓后命令机器人抓握物体,并把它运送到指定的地点从而达到物体分类的目的。

图象噪声的模糊聚类平滑方法

模糊C-means算法在聚类分析中已得到了成功的应用,本文提出一种利用模糊C-means算法消除噪声的新方法。一般来说,图象中的噪声点就是其灰度值与其周围象素的灰度值之差超过某个门限值的点。根据这个事实,首先利用模糊C-means算法分类,再利用标准核函数检测出噪声点,然后将噪声点去掉。由于只修改噪声点处的象素灰度值,而对于其它象素的灰度值不予改变,所以本算法能够很好地保护细节和边缘。本方法每次处理3×3个点,而以往的方法只能每次处理一个点,所以本方法能提高运算速度。文中给出了利用本方法对实际图象处理的结果,并与梯度倒数权值法进行了定量的比较。

嵌入式智能图像处理在火灾预警系统的应用

传统的火灾检测方法一般采用感烟、感温、感光探测器等进行探测。本文提出了一种嵌入式基于图像视觉特征的火灾检测方法,以TI公司的数字多媒体处理器TMS320DM642为核心,设计实现智能前端火灾探测与自动报警系统。通过DM642对视频图像进行采集并结合相应的智能图像处理与模式识别算法,对森林火险进行实时监控。实验结果表明,该系统比传统系统更进一步减少了误报率且具有响应快、监控范围广等优点。

ETC中具有车型识别功能的车辆检测器设计

介绍了专门用于ETC(不停车收费系统)中一种车辆检测器的软硬件设计方法。根据车辆检测器应用环境的特点给出了基准频率校正算法,可以对基准频率进行实时校正。并采用模糊模式识别算法进行车型识别。

基于模糊模式识别的车型分类研究

根据目前中国路桥车辆收费标准,提出了一种基于模糊模式识别的车型分类系统。车辆经过环形线圈传感器时,形成感应曲线,提取感应曲线的特征并进行特征分离,利用模糊模式识别方法对车型进行匹配分类。研究结果已在路桥收费系统以及交通流量统计中得到应用。

断陷湖盆缓坡带成藏模式研究

Halfgraben-like depressions have multiple layers of subtle traps, multiple coverings of oil-bearing series and multiple types of reservoirs. But these reservoirs have features of strong concealment and are difficult to explore. For this reason, many scholars contribute efforts to study the pool-forming mechanism for this kind of basins, and establish the basis for reservoir exploration and development. However, further study is needed. This paper takes HuiMin depression as an example to study the pool-forming model for the gentle slope belts of fault-depression lake basins. Applying multi-discipline theory, methods and technologies including sedimentary geology, structural geology, log geology, seismic geology, rock mechanics and fluid mechanics, and furthermore applying the dynamo-static data of oil reservoir and computer means in maximum limitation, this paper, qualitatively and quantitatively studies the depositional system, structural framework, structural evolution, structural lithofacies and tectonic stress field, as well as fluid potential field, sealing and opening properties of controlling-oil faults and reservoir prediction, finally presents a pool-forming model, and develops a series of methods and technologies suited to the reservoir prediction of the gentle slope belt. The results obtained in this paper richen the pool-forming theory of a complex oil-gas accumulative area in the gentle slope belt of a continental fault-depression basin. The research work begins with the study of geometric shape of fracture system, then the structural form, activity stages and time-space juxtaposition of faults with different level and different quality are investigated. On the basis of study of the burial history, subsidence history and structural evolution history, this paper synthesizes the studied results of deposition system, analyses the structural lithofacies of the gentle slope belt in the HuiMing Depression and its controlling roles to oil reservoir in the different structural lithofacies belts in time-space, and presents their evolution patterns. The study of structural stress field and fluid potential field indicates that the stress field has a great change from the Dong Ying stages to nowadays. One marked point among them is that the Dong Ying double peak- shaped nose structures usually were the favorable directional area for oil and gas migration, while the QuDi horst became favorable directional area since the GuanTao stage. Based on the active regular of fractures and the information of crude oil saturation pressure, this paper firstly demonstrates that the pool-forming stages of the LingNan field were prior to the stages of the QuDi field, whici provides new eyereach and thinking for hydrocarbon exploration in the gentle slope belt. The BeiQiao-RenFeng buried hill belt is a high value area with the maximum stress values from beginning to end, thus it is a favorable directional area for oil and gas migration. The opening and sealing properties of fractures are studied. The results obtained demonstrate their difference in the hydrocarbon pool formation. The seal abilities relate not only with the quality, direction and scale of normal stress, with the interface between the rocks of two sides of a fault and with the shale smear factor (SSF), but they relate also with the juxtaposition of fault motion stage and hydrocarbon migration. In the HuiMin gentle slope belt, the fault seal has difference both in different stages, and in different location and depth in the same stage. The seal extent also displays much difference. Therefore, the fault seal has time-space difference. On the basis of study of fault seal history, together with the obtained achievement of structural stress field and fluid potential field, it is discovered that for the pool-forming process of oil and gas in the studied area the fault seal of nowadays is better than that of the Ed and Ng stages, it plays an important role to determine the oil column height and hydrocarbon preservation. However, the fault seal of the Ed and Ng stages has an important influence for the distribution state of oil and gas. Because the influential parameters are complicated and undefined, we adopt SSF in the research work. It well reflects synthetic effect of each parameter which influences fault seal. On the basis of the above studies, three systems of hydrocarbon migration and accumulation, as well as a pool-forming model are established for the gentle slope belt of the HuiMin depression, which can be applied for the prediction of regular patterns of oil-gas migration. Under guidance of the pool-forming geological model for the HuiMin slope belt, and taking seismic facies technology, log constraint evolution technology, pattern recognition of multiple parameter reservoir and discrimination technology of oil-bearing ability, this paper develops a set of methods and technologies suited to oil reservoir prediction of the gentle slope belt. Good economic benefit has been obtained.

A Theory of Networks for Appxoimation and Learning

Learning an input-output mapping from a set of examples, of the type that many neural networks have been constructed to perform, can be regarded as synthesizing an approximation of a multi-dimensional function, that is solving the problem of hypersurface reconstruction. From this point of view, this form of learning is closely related to classical approximation techniques, such as generalized splines and regularization theory. This paper considers the problems of an exact representation and, in more detail, of the approximation of linear and nolinear mappings in terms of simpler functions of fewer variables. Kolmogorov's theorem concerning the representation of functions of several variables in terms of functions of one variable turns out to be almost irrelevant in the context of networks for learning. We develop a theoretical framework for approximation based on regularization techniques that leads to a class of three-layer networks that we call Generalized Radial Basis Functions (GRBF), since they are mathematically related to the well-known Radial Basis Functions, mainly used for strict interpolation tasks. GRBF networks are not only equivalent to generalized splines, but are also closely related to pattern recognition methods such as Parzen windows and potential functions and to several neural network algorithms, such as Kanerva's associative memory, backpropagation and Kohonen's topology preserving map. They also have an interesting interpretation in terms of prototypes that are synthesized and optimally combined during the learning stage. The paper introduces several extensions and applications of the technique and discusses intriguing analogies with neurobiological data.

Template Matching: Matched Spatial Filters and Beyond

Template matching by means of cross-correlation is common practice in pattern recognition. However, its sensitivity to deformations of the pattern and the broad and unsharp peaks it produces are significant drawbacks. This paper reviews some results on how these shortcomings can be removed. Several techniques (Matched Spatial Filters, Synthetic Discriminant Functions, Principal Components Projections and Reconstruction Residuals) are reviewed and compared on a common task: locating eyes in a database of faces. New variants are also proposed and compared: least squares Discriminant Functions and the combined use of projections on eigenfunctions and the corresponding reconstruction residuals. Finally, approximation networks are introduced in an attempt to improve filter design by the introduction of nonlinearity.

Fuzzy-Rough Sets Assisted Attribute Selection

R. Jensen and Q. Shen. Fuzzy-Rough Sets Assisted Attribute Selection. IEEE Transactions on Fuzzy Systems, vol. 15, no. 1, pp. 73-89, 2007.

Semantics-Preserving Dimensionality Reduction: Rough and Fuzzy-Rough-Based Approaches

R. Jensen and Q. Shen. Semantics-Preserving Dimensionality Reduction: Rough and Fuzzy-Rough Based Approaches. IEEE Transactions on Knowledge and Data Engineering, 16(12): 1457-1471. 2004.

Feature Selection based on Rough Sets and Particle Swarm Optimization

X. Wang, J. Yang, X. Teng, W. Xia, and R. Jensen. Feature Selection based on Rough Sets and Particle Swarm Optimization. Pattern Recognition Letters, vol. 28, no. 4, pp. 459-471, 2007.

Rough sets, their extensions and applications

Q. Shen and R. Jensen, 'Rough sets, their extensions and applications,' International Journal of Automation and Computing (IJAC), vol. 4, no. 3, pp. 217-218, 2007.

Fuzzy-Rough Data Reduction with Ant Colony Optimization.

R. Jensen and Q. Shen, 'Fuzzy-Rough Data Reduction with Ant Colony Optimization,' Fuzzy Sets and Systems, vol. 149, no. 1, pp. 5-20, 2005.

Selecting Informative Features with Fuzzy-Rough Sets and its Application for Complex Systems Monitoring.

Q. Shen and R. Jensen, 'Selecting Informative Features with Fuzzy-Rough Sets and its Application for Complex Systems Monitoring,' Pattern Recognition, vol. 37, no. 7, pp. 1351-1363, 2004.