基于改进SVM分类器的动作识别方法

利用智能手机加速度传感器信号,提出一种改进的动作识别方法以降低传统动作识别方法的复杂程度,提高识别率。在特征提取时用盲选法,即用PCA(principal component analysis)进行特征值的降维和去除多维间的干扰,而所选特征没有对应的物理意义;并在分类识别中将遗传算法应用到SVM(support vector machine)分类器参数优化中。通过实验表明,该方法能够对日常的走路、站立、跑及上下楼等动作进行准确的识别。     

A Self-Learing Chinese Character Recognition System

A method using possibility theory and fuzzy logic is introduced. Possibility functions of each feature are determined during training. Methods of fuzzy logic are used to combine the possibility of different features. This Chinese character recognition system consists of a pre-classifier and a two-level fine classifier. In order to handle deviations of the characters,a fuzzification of features is introduced. This method is used to develop single-font classifiers. The whole system consists of a font-classifiers and four single-font-classifiers.     

基于移动溯源与图像分析的茶叶品级鉴定方法研究

为解决茶叶在质量追溯时防伪和品质鉴定问题, 以六安瓜片为研究对象, 提出一种结合移动计算和图像分析的茶叶质量追溯方法。该方法的实现以 Android操作系统为平台, 通过扫描茶叶质量追溯二维码查询溯源信息。在此基础上, 利用智能终端拍照获取完整的单片茶叶正面照片并上传至后台服务器。后台编写的 MATLAB图像处理程序对采集的图像进行预处理和特征参数提取, 将提取的颜色和形状共 7项特征参数作为参数集合, 运用 SVM分类器对茶叶进行分类, 最终将结果反馈到溯源客户端。测试结果证明, 该方法能快速帮助消费者对茶叶品级进行初步鉴定。     

The Diagnostic Model of Cotton Aphids Based on Leaf Textural Features

[Objective] The aim of this research is to explore the possibility of rapid identification of aphid damage to cotton leaves by identifying the hyperspectral image about healthy and aphid damage cotton leaves. [Method] Taking Xinluzao 45 as material, hyperspectral images of healthy and damaged cotton leaves by aphids were obtained, and hyperspectral image information was extracted from interested regions of different cotton samples. Then three descending dimension methods were used to extract hyperspectral feature, and Gray-Level Co-occurrence Matrix to extract image texture feature. Finally a cotton aphids damage diagnostic model was built up. [Result] The prediction accuracy based on all textural samples input Random frog-partial least-square-linear discriminant Function (RF-PLS-LDA) model was 91.49%. The prediction accuracy based on energy data input principal component analysis-loading partial least-square-linear discriminant function (PCA-Loading-PLS-LDA) model was 92.55%. [Conclusion] The second-order statistics (energy) of gray co-occurrence matrix can be used to simplify the model, reduce the computation and improve the stability of prediction. Based on the texture feature vector, the identification of aphid cotton leaves can be realized effectively, which provides a method for the rapid identification of insect pests.     

基于颜色和纹理特征的花生仁外观品质检测研究

为了提高对花生仁外观缺陷的在线分类准确率及效率。通过对采集完好、破损、霉变的花生仁RGB图像进行均值位移法、灰度处理以及阈值分割等预处理,研究提取了花生仁HSV颜色空间下的H、S、V各分量的一阶矩和二阶矩共6个颜色特征值,再基于灰度共生矩阵法提取能量、熵、对比度、逆差分矩共4个纹理特征值,构建颜色和纹理结合的特征向量,最后分别采用BP神经网络和SVM分类器对花生仁进行分类识别。结果表明：在花生仁的整体识别准确率上,BP神经网络为96.67%,SVM分类器为97.22%,后者优于前者,在识别时间上BP和SVM分别为2.5 s和1.1 s,识别效率上也是SVM更好,综合识别准确率和效率两方面考虑,优先选择SVM分类器模型来对花生仁进行分类识别。     

Application of Gobor wavelet and SLLE in face recognition

In order to improve the recognition rate of face recognition algorithm, a new algorithm of face recognition is proposed based on Gabor wavelet transform and Supervised Locally Linear Embedding (SLLE). Gabor wavelet is introduced as a method to extract Gabor magnitude features by convolving the normalized face image with multi scale and multi orientation Gabor filters. In the feature extraction module, the dimension of Gabor features is reduced by SLLE. A minimum distance classifier is trained for classification. With the test of the ORL and YALE face database, it is found that 3.5 %~37.8% increase in recognition rate can be achieved compared with other algorithms.     

面向植物生长智能监控的叶颜色特征提取

根据植物生长智能监控的需求,设计了叶颜色特征提取方法：首先提取叶整体颜色特征,根据特征数据能够判断植物叶是健康、存在整体颜色变异还是存在局部颜色变异;如果存在局部颜色变异,进行局部颜色变异部分分割、颜色特征提取和空间分布特征提取。将一组具有不同程度颜色变异的植物叶的阳面和阴面图像作为两个系列进行了对比测试。测试数据表明：特征参数能够区分健康的叶和存在变异的叶,能够区分不同类型和不同程度的颜色变异,能够为后续决策处理提供数据支持,特征提取方法适合植物生长智能监控。     

Pd Pattern Recognition Based on Linear Discriminant Analysis in GIS

According to the character of PD in GIS, the authors design four kinds of GIS defection models. The GIS gray intensity images are constructed based on mass specimens gathered by the ultra - high frequency and high speeds systems, Aiming at the PD characteristics and its defections, A PCA-FDA method is put forward based on PD images. The principal component analysis is employed to condense the dimension of PD images, then the optimal sets of statistically uncorrelated discriminant vectors are extracted, and the minimum distance classifier is constructed as classifier. The identified results show that this method can effectively elevated the discrimination of the four kinds of defects in GIS PD.     

Feature Extraction and the Design of Classifier in Small Set Handwritten Chinese Character Recognition

This paper presents basic concept of feature extraction of handwritten Chinese character, and proposes a new feature extraction named superposition mesh weighting factors strokes extracting algorithm to obtain feature of small set handwritten Chinese character. Basing on the analysis model of RBFNN, an integration RBF classifier is used for small set handwritten Chinese character recognition. Then, the hybrid optimize strategy, which combines the genetic algorithm and the simulated annealing, is adopted to train RBFNN.     

Genetic diversity of high-molecular-weight glutenin subunit compositions in landraces of hexaploid wheat from Japan

The high-molecular-weight (HMW) glutenin subunit composition of seed storage protein of 174 Japanese hexaploid wheat (Triticum aestivum) landraces have been examined by using sodium dodecyl sulfate polyacrylamide gel electrophoresis system. Twenty four different, major glutenin HMW subunits were identified, and each of the landraces contained three to five subunits and 17 different glutenin subunit patterns were observed for 13 alleles in the landraces. On the basis of HMW glutenin subunits composition, Japanese landraces showed a specific allelic variation, close to Japanese commercial wheats in HMW glutenin subunits, different from those in alien hexaploid wheats. Further, it could be concluded that all common glutenin alleles can be found in the 174 landraces originated from Japan. The variation detected in the glutenin subunits is useful for variety identification, has a bearing on our understanding of hexaploid wheat genetic resource evolution in Japan, and raises questions concerning the nature of this genetic variation. This revised version was published online in July 2006 with corrections to the Cover Date.     

A data transfer method for improving seed identification of maize (Zea mays) haploid breeding based on genetic similarity

Maize haploid breeding technology is able to identify haploid seeds non‐destructively, rapidly and at low cost with the help of Near‐infrared (NIR) spectral analysis. However, due to the hybridization of numerous parents and the low production rate of haploid, the haploid data collection becomes a burden for engineering this technology. Biologically, there are considerable similarities between the progeny of the same female parent and different male parents. Based on this advantage, similar spectral data can be transferred when the NIR technology is employed. A revised method of Transfer adaptive boost (TrAdaBoost) is proposed to improve identifying for the backpropagation neural network (BPNN) classifier. To avoid the negative transfer, a screening thresh is used to select out similar data, and the amount of these data are limited to join current training. The results show that the identification performances are improved significantly when the data amount is small. This method shows a high ability to make the seed identification more convenient for engineering maize haploid breeding.     

植物叶裂特征自动提取研究

为了实现叶裂特征自动提取,构造了植物叶片凸包距离函数,通过分析该函数,判断是否存在叶裂,如果存在叶裂,进一步判别叶裂类型,获取叶裂数量与深度数据。利用树叶数据库中的图像对方案进行了测试,结果显示：叶裂存在性参数对叶片是否具有叶裂区分明显,叶裂位置检测、叶裂数目判定准确,叶裂类型判别参数体现了不同叶裂类型的差异,叶裂深度参数体现了叶裂的深度差异。测试表明方案切实可行,可用于植物的种类自动识别和长势监控系统。     

Identification of different aromatic Chinese liquors by colorimetric artificial nose

A potential instrument, the colorimetric nose, is developed to identify the fragrances of five different Chinese liquors. Firstly, in order to minimize error, the RGB values of the raw data are preprocessed using a threshold before further analysis. The output of the artificial nose is then analyzed by hierarchical routing cluster analysis (HCA), principal component analysis (PCA) and support vector machine (SVM). It is found that HCA can perform correctly classify fragrances into five different classes. However, using the first three components identified by PCA analysis, representing 80.79% of the variance, the five individual fragrances can be reliably distinguished. Finally, it also shows that the five constituent fragrant liquors can also be reliably classified with 100% accuracy by SVM. These results show that the colorimetric artificial nose, a simple and efficient detection and identification tool, has great potential to identify different constituent fragrant liquors reliably well.     

Identifizierung verschiedener Miscanthus-Sorten mittels Isoenzymen

Identification of Miscanthus Varieties by their Isozymes
With the aim of testing the possibility of identifying Miscanthus varieties, zymograms of 13 isozymes from 65 plants of M. smensis, M. × giganteus and M. sacchariflorus were analysed using starch-gel electrophoresis. The zymograms of root, rhizome, and leaf tissues were compared. After similarity between the zymograms of the different types of tissues had been established, the tissue of young leaves was used for analysis. The zymograms of the different isozymes are variable in most cases and show as many as nine patterns. The large variation between the varieties allows an identification of each of the varieties analysed. About half of the plants could be attributed to M. × giganteus. In this group of plants, different variants were found in the isozymes of diaphorases and glutamate-oxalacetate-transammases. These variants might indicate that M. × giganteus includes more than one genotype. Some of the plants which were assumed to belong to M. × giganteus could clearly be excluded from this variety. Although some of the varieties can only be unequivocally identified by their flowers, it is possible to analyse the isozymes even in the dormant state of the rhizomes, thus, this method offers a simple way of identifymg Miscanthus varieties by the principle of exclusion, independent of the physiological state of the plant.     

Anomaly detection based on the multi phase clustering and naive bayes

Anomaly detection method was used for calibration data concentration significantly different from other data objects. In this paper, the multi phase clustering aimed at resolving the import of noise data and the lack of the attributive sample, and changing the traditional passive learning of bayes for active learning ways to structure the superior performance classifier. In the pre processing stage, a clustering algorithm based on density is introduced to handle noise data. And the output of the density based clustering algorithm can be used as the input of K means, which responsible for handling the training samples with absent values. At classification time, we introduce adaboost algorithm into naive bayes to generate a more effective classifier.     

Face recognition method with local matching

From human cognition, a face recognition method with local matching based on statistical learning is proposed. The image is divided into several subimages and each subimage is considered as a weak classifier. The Adaboost learning algorithm is used to train the weak classifiers and construct a strong classifier. As a result, all subimages are effectively combined together to explore the best discriminating power and improve the classification accuracy. Compared with the holistic matching methods, the local matching method is robust to variations in illumination, expression, and pose, etc. The experimental results show that the proposed method can improve the face recognition accuracy and is robust to variations in illumination and expression.     

高效液相色谱法在大白菜种子品种鉴定中的应用

该文用高效液相色谱法对北京普遍种植的12个大白菜品种种子的水溶性蛋白进行了色谱分离,同时对北京新二号及其亲本的种子也进行了分析比较。根据不同品种由大小不同的主要色谱峰组成的“指纹”图谱,各品种可以很容易得以鉴别。结果表明:作者发展的高效液相色谱法很适于大白菜种子的品种鉴定.并且为大白菜育种中数量性状的遗传分析提供了有力的手段。     

Comparative Research of Six Kinds of Callicarpa Medicinal Herbs in the Character Identification,Semi-microscopic Identification and Microscopic Identification

[Objective]To establish a simple and easy identification method for six kinds of Callicarpa medicinal herbs,to ensure the clinical medication safety,and to provide references for the quality control of Callicarpa medicinal herbs. [Methods] Character identification,semimicroscopic identification and microscopic identification were used to identify the six kinds of Callicarpa medicinal herbs. [Results]In external characters,they were identified mainly through color,hair and morphological characteristics. In microscopical characters,they were identified though the form,distribution and quantity of glandular hairs and nonglandular hair; crystal types in leaves were used for supplementary identification of Callicarpa medicinal herbs. [Conclusions] Characters identification method could identify several Callicarpa medicinal herbs,while some varieties were difficult to distinguish. Different kinds of Callicarpa could be identified by using semi-microscopic identification and microscopic identification methods to observe the morphological features and distribution characteristics of glandular hairs,non-glandular hairs and crystals. An accurate,rapid and economical microscopic identification method was important to identify leaves of different varieties of Callicarpa.     

Modal parameters identification for short data sequences based on stratified sampling and optimism complex Morlet wavelet

A novel modal parameter identification method based on stratified sampling and optimism complex Morlet wavelet is proposed for short data sequences. Stratified sampling is applied to divide the structure response signal into different layers which called sub samples with different thresholds, and then free decrement response signal of each layer is extracted by random decrement technique. The optimism complex Morlet wavelet transform is applied to identify modal parameter of each layer, and the weight of the layer is also determined based on the sample standard deviation. The modal parameter of the structure can be obtained by weighted calculation.The engineering application shows that the proposed method has the ability to identify modal parameter accurately, decouple low frequency intensive modal composition and restrain high frequency fake modal effectively.