基于动平台多源数据的运动目标测量方法研究

根据运动目标性能试验需求及试验平台的特点,建立一个影像数据测量系统,实现目标的运动视频记录和外部参数测量。采用基于OpenCV的棋盘格内参数标定结合基于平台上固定标识点、线为基准的外场标定的摄像机标定方法,采用高速影像数据、中线视频数据和GPS数据相结合的方法实现目标运动的轨迹、姿态等参数的测量。通过测量精度分析及多次试验结果证明,测量精度达到水平方向优于10 cm,高程方向优于5 cm,姿态精度优于05。     

A recursive color image edge detection method using Green's function approach

In this paper, an extended version of image edge detector using Green's function approach is proposed for detection of edges in the color vector space field. In the proposed method, the relationship between the Red, Green and Blue components is considered to design a differential operator for detection of edges in color images. By using the proposed operator, partial derivatives of all components of color image can simultaneously affect on the edge detection process. Therefore the proposed method can preserve the vector nature of color images during the edge processing stages. Also, the proposed method is compared both quantitatively and qualitatively with other color edge detectors. Experimental results show that the proposed method can efficiently preserve the edges even when the color images corrupted with different levels of noise.     

Fingerprint image enhancement using data driven Directional Filter Bank

A new method based upon data driven tool, principal component analysis (PCA), for fingerprint enhancement is proposed in this paper. PCA is a very useful statistical technique that has found application in many different fields like image compression, face recognition and is commonly used for finding patterns in data of high dimension. In the proposed method, the input image is first decomposed into directional images using decimation free Directional Filter Bank (DDFB). Then these directional images are normalized. A data driven technique PCA is applied to these normalized directional fingerprint images, which gives the PCA filtered images. These are basically directional images. Then these directional images are reconstructed into one image which is the enhanced one. Simulation results are included illustrating the capability of the proposed method.     

Cryptanalysis and improvement of a digital image encryption method with chaotic map lattices

A digital image encryption scheme using chaotic map lattices has been proposed recently. In this paper, two fatal flaws of the cryptosystem are pointed out. According to these two drawbacks, cryptanalysts could recover the plaintext by applying the chosen plaintext attack. Therefore, the proposed cryptosystem is not secure enough to be used in the image transmission system. Experimental results show the feasibility of the attack. As a result, we make some improvements to the encryption scheme, which can completely resist our chosen plaintext attack.     

基于高光谱的脱绒棉种电导率快速检测研究

为了寻求一种快速、无损检测脱绒棉种活力的方法,提出基于高光谱技术预测脱绒棉种电导率。采集了新陆早50、新陆早57、新陆早62三个品种且不同老化程度下共810粒脱绒棉种高光谱图像(400~1 000 nm),通过组合不同预处理方法,采用chauvenet检测方法剔除异常值后建立了偏最小二乘法(PLS)、逐步多元线性回归(SMLR)、主成分回归(PCR)模型。结果表明,采用变量标准化(SNV)、卷积平滑(Savitzky-Golay)、一阶微分(First derivative)和norris微分平滑组合的预处理方法,波段范围为480~530,650~980 nm下建立的PLS模型效果最佳;其中PLS模型得到新陆早50、新陆早57、新陆早62的预测集相关系数和校正集相关系数分别为0.88,0.90,0.92,0.91,0.89,0.90;预测集均方根误差(RMSEP)和校正集均方根误差(RMSEC)分别为44.3,38.4,37.8,46.5,43.5和40.8 μS·cm-1。研究结果表明,采用高光谱技术预测脱绒棉种电导率具有一定的可行性,也为其他种子的活力检测奠定了良好的基础。     

高光谱成像的油菜和杂草分类方法

利用高光谱成像技术结合化学计量学方法对油菜中的杂草进行分类识别。采用近红外高光谱技术,通过正态变量变换(SNV)、去趋势化(De-trending)、多元散射校正(MSC)、移动平均平滑法(MA)、多项式卷积平滑法(SG)、基线校正(baseline)及归一化(normalize)算法对光谱数据进行预处理,采用主成分载荷(PCA loadings)、载荷系数法(x-LW)、回归系数法(RC)、连续投影算法(SPA)分别进行特征波长提取,采用偏最小二乘判别分析(PLS-DA)、极限学习机(ELM)和支持向量机(SVM)建立分类模型。结果表明,基于De-trending 预处理,通过PCA loadings,x-loading weights及SPA特征波长提取方法,基于极限学习机ELM算法建立的模型取得了最优的分类效果,建模集和预测集的分类精度均达到100%,另引入平均分类精度的指标,发现不同试验时间下,模型分类精度变化不大,表明应用近红外高光谱成像技术对油菜和杂草进行分类是可行的。     

球粉板表面缺陷在线检测

为快速准确检测球粉板表面缺陷,对球粉板表面光学特性不一致性和缺陷类型的多样性等关键问题进行了分析,以利于有效地解决关键问题;论文提出采用基于模板图片创建矫正模板来解决被测板材表面光学特性变动的均化问题、自适应萃取二值化阈值解决二值化阈值整定问题,以提高软件的执行鲁棒性;通过自定义算法实现了细线型缺陷修补,有效地提高了划痕类缺陷检测的精度和可靠性;提出了筛查模型实现了符合缺陷检测精度要求的噪点筛除;通过大量现场实验验证了本系统能够正确高效实现球粉板表面缺陷检测定位标识和面积测量功能,以及达到了目标检测精度;本系统能很好地胜任球粉板在线表面缺陷检测;实验证明检测系统非常高效精准,极大地提高了生产线自动化能力。     

红外图像的疲劳状态识别方法

目的：为了解决光照变化对疲劳检测系统造成的准确性不高的问题,提出了一种近红外环境下判断人眼状态的方法,即针对红外光补图像的人眼状态判断。方法：首先,利用Adaboost算法对人眼区域进行定位,在网格法标记人眼瞳孔部分的基础上,进行Retinex红外图像增强。接着,对二值化与边缘检测后的红外图像分别进行网格法闭合度计算,得到人眼闭合度。最后,根据闭合度计算结果设定双阈值并结合PERCLOS来判断眼睛状态。结果：在DM642硬件平台上进行疲劳检测试验,实验结果表明,人眼状态识别率达到了90%以上,且平均每秒能处理21帧图片。结论：证明了该方法不仅能有效解决光照变化带来的问题,而且满足疲劳状态检测系统的快速性、可靠性和有效性等要求。     

基于图像块能量梯度的多聚焦图像融合

针对多聚焦图像,提出一种基于图像分块的融合方法。将源图像分为大小相同数量相等的子块,采用能量梯度算子作为对焦评价函数,计算各个图像子块能量梯度匹配度,设置匹配度阈值分离出源图像中的清晰区域。源图像中的清晰区域直接作为融合图像相应的区域,其它区域的处理中,构造与相应子块能量梯度大小相关的图像序列,以及像素点到各个子块中心距离相关的融合函数,然后用融合函数对图像序列融合。实验结果表明该方法有效性和合理性。     

基于VxWorks新型映像的三模冗余启动机制研究

针对VxWorks启动映像依赖于ROM地址,不能适应航天型号中三模存储要求的缺点,利用VxWorks自带的压缩算法,按照航天型号软件任务要求设计了一种基于VxWorks内存型映像的新型压缩映像,实现了软件映像与ROM区地址无关。并且实现了新型压缩映像的三模冗余引导启动,启动时每一个bit数据都从三份映像获取,并进行三取二判断,将最后结果拷贝到内存作为目的映像的bit数据,目的映像启动后能够回写正确数据自动纠正空间环境下可能发生的单粒子翻转错误,为软件的可靠启动建立了基础。