基于二维成像的三维物体形状特征分析

目前三维物体识别方法在识别过程中所需数据量大、难以实用。该文简化三维物体识别过程,构建了一个基于单视点二维投影图像的三维物体识别系统。分别选取Zernike矩、基于Trace变换的Triple特征、MSA等三种形状特征,实现了对物体的视点空间的聚类划分。在普林斯顿三维模型库上,通过分类识别实验分析三种形状特征的性能。实验表明:特征对不同类别物体的分类效果差异明显。该文由此提出了针对目标物体形状及应用环境的特征选取方案。

茄科类作物杂草的形状特征研究

杂草影响作物生长,除草剂的大面积使用既增加了农民的经济负担,又污染了环境。根据茄科类作物及其杂草的形状特点,提取其有效的形状特征参数,并用神经网络进行验证。试验证明,提取的形状特征能有效地识别出杂草。

一种自动的高分辨率遥感影像道路提取方法

从高分辨率遥感影像中提取道路有着非常重要的意义,但是受到遥感影像噪声、复杂的自然场景和已有算法的局限性的影响,道路提取有待于进一步研究.近些年来水平集方法被用于提取道路,但是初始水平演化曲线的确定却是一个大的难点.笔者提出一种自动的水平集分割方法,并将其用于道路检测中.首先,将卷积神经网络用于道路的粗分类.然后,利用形状特征和孔洞填充方法得到比较准确的道路区域.在此基础上,利用张量投票来提取道路的交叉口,并将其轮廓作为水平集演化的初始曲线进行水平集分割.最后,结合卷积神经网络分类和水平集分割的优势,得到比较完整的道路区域,并保持了道路的边缘.实验结果表明,该方法能自动地提取准确完整的道路区域.

一种高分辨率遥感图像道路自动提取方法

高分辨率遥感图像道路提取是遥感界关注的重要研究领域,如何快速、有效、智能地进行遥感信息分析和处理是当今遥感界迫切需要解决的问题,自动道路提取方法称为道路提取研究的热点。针对道路提取受遥感图像噪声和树木阴影等复杂自然场景因素影响的问题,提出一种基于卷积神经网络(CNN)和数学形态学算法相结合的自动道路提取方法。首先,构建出深度卷积神经网络模型,在训练卷积神经网络模型时,通过改进的批量随机梯度下降算法(MBGD)来训练深度卷积神经网络;然后,利用训练好的深度卷积神经网络进行道路特征的提取;最后,结合数学形态学优化算法进一步优化道路提取结果。实验结果表明,该方法能提取出完整的道路区域。

Rules of distribution in a plastic zone of rocks surrounding a roadway affected by tectonic stress

In order to study the rules of distribution in a plastic zone of rocks, surrounding a roadway, affected by tectonic stress, we first analyzed the mechanics of a roadway affected by tectonic stress and derived a theoretical formula for the plastic zone of rocks surrounding a roadway. We also analyzed the distribution characteristics of the plastic zone under different levels of tectonic stress, vertical pressure, cohesion and friction angle of the surrounding rock. Secondly, we used numerical simulation to analyze the range and shape features of the plastic zone of rocks surrounding the roadway, given different tectonic stress levels. Finally we used a rock drilling detector to carry out field measurements on the broken state of rock surrounding the roadway at the –700 substation and channels in the Xinzhuang mine of the Shenhuo mining area. Given the measured ground stress, we analyzed the relationship between tectonic stress and the distribution of this plastic zone. Our results show that the range of the plastic zone at the top and bottom of the roadway increases with an increase in tectonic stress and this increase is especially obvious at the roadway corner.

The shape character and development stage of nebkha

Based on statistical analysis on morphological data of nebkhas surveyed in the Hotan River Basin,Xinjiang, the changing regularity and characters of nebkhas’ morphology in different development stageshave been discussed. Results indicate that there exist better correlation among length, width and altitudeof all kinds of nebkhas, but there are certain various shape characters in different nebkhas and the sametype of nebkhas in differential development stages. On account of analysis on mathematical imitation ofnebkhas’ shape characters, regional ecological environmental and aeolian characters, the development ofthe nebkha can be divided into a growing stage, a stabilizing stage and a declining stage.

Morphological Study of Ficus deltoidea Jack in Malaysia

Ficus deltoidea Jack （Moraceae） or Mas Cotek is a small shrub with a great morphological variation. Measurement of 40 morphological traits had been done on 50 accessions to find the most significant characters that enable differentiation being done according to its variety groups. The data were analyzed with principal component analysis （PCA） and cluster analysis （CA） using cluster software package programme to produce the scatter diagram and dendrogram relationship of the taxa. The results showed that there were 25 morphological characters having the value of factor analysis greater than 0.60 from its principal component （PC） with the Eigen value greater than 1.0. 16 out of 40 morphological characters had been identified as having high values of correlation coefficient ranging from -0.783 to 0.890. The analysis showed that the most responsible characters in grouping the samples into different groups are the shape and size of leaf, number and color of dots on the leaf surface and characters of syconium. The scatter diagram of the accessions on the PC1 against PC2 showed six major groups. The dendrogram displayed the relationship among the accessions and within the dissimilarity distance = 19, it classified the samples into five major groups. Observation on F. deltoidea resulted in the findings of high variability among the accessions. The most significant characters in grouping accessions are the shapes of leaf base （BL）, shape of leaf apex （SA）, ratio of lamina width to lamina length （R）, dots color at the lower midrib （DLM）, color of young syconium （CYS）, color of mature syconium （CMS） and the number of syconium on trees （DST）. This study provides basic information for introduction of some particular traits and effective conservation of the species breeding programme. The morphological traits were found to be useful for the diversity studies and in identifying the variation. The actual figures of F. deltoidea obtained through this study enable comparison being done to the previous and in future study. Hence, the varieties that are extinct could be recognised.

Method for extracting geometrical characteristics of joint probability density based on contour lines

This paper presents a method for extracting geometrical features of the joint probability density function(PDF)of two-dimensional systems based on its contour lines,with particular interests given to the number and position of peaks and craters.In order to detect those two types of structures,a series of horizontal planes are applied to truncate the joint PDF with contour lines generated.Starting with the analysis of contour lines in a single plane,shape characteristics of the peak and the crater can be reflected on the contour lines in the aspects of gradient direction and inclusion relationship.Aided by the properties of PDF,the information about gradient direction and inclusion relationship of contour lines can be obtained simultaneously if the contour tree is built.According to the contour tree,the contour lines can be classified as two groups.Then the corresponding relation between contour lines in different planes is discussed.Based on the corresponding relation,clustering analysis about contour lines belonging to the same group but having different heights is performed.Two sets of contour lines are finally obtained as the simplest expression of geometrical characteristics of a joint PDF.They can be used to obtain the number and position of each peak and crater.Three oscillators of different types are chosen to test this method,which shows that this method can pave the way for numerical calculation about the stochastic P-bifurcation of multi-dimensional systems.