花生联合收获机对辊差相半喂入摘果装置的研发

针对花生联合收获机现有的摘果装置摘净率及纯生产率低、破损率和含杂率高、机械结构不稳定、故障率高、结构复杂制作成本高等问题,设计了半喂入对辊差相摘果装置。并对摘果装置进行结构设计与性能研究,确定该装置的最优结构和组合参数。该研究可大大提高花生联合收获机的收获效率,降低花生收获成本。     

基于函数型数据时间序列建模的齿轮破损图像识别

当前齿轮破损图像识别未拟合周期性动作捕捉数据,导致齿轮破损图像的识别效率较低,图像边缘轮廓中细节信息丢失,造成识别误差较大、识别效果较差。提出基于函数型数据时间序列建模的齿轮破损图像识别方法,通过小波变换方法提取齿轮破损图像中存在的低频成分,做双直方图均匀化处理,采用高斯高通滤波器提取齿轮破损图像中存在的高频成分,增强齿轮破损图像中存在的细节信息,利用函数型数据分析方法拟合动作捕捉数据,根据获取的数据构建周期时间序列,获得隐马尔可夫模型,通过最大似然估计函数,计算获得测试模型与样本之间的匹配度,实现齿轮破损图像的识别。实验结果表明,所提方法的识别误差较小,识别效果较好,能够有效提高齿轮破损图像的识别效率。     

基于多模态感知融合的钛合金加工刀具磨破损分析

飞机钛合金加工铣刀磨破损量直接影响加工质量与铣刀使用寿命,然而现有方法难以实现在线直接有效测量铣刀的磨破损量,给飞机制造带来较大的质量隐患。为实现铣刀磨破损的直接有效在线测量,文中提出了识别测量算法,主要包含基于卷积神经网络识别并获取有效模态特征图像,多模态感知融合的刀具磨破损分析策略识别破损区域,基于多模态融合分析得到具体的刀刃磨破损量化数值,并根据融合结果实现刀具的磨破损评价。对4把钛合金加工铣刀进行实验,提出的方法可准确识别出异常图像及区域,同时max-标准偏差分别仅占标准值的1.66%、3.52%、2.57%、2.04%。结果表明,说提出方法良好的磨破量感知融合分析能力,结合装置可实现在线准确测量磨破损量,为工程应用奠定了基础。     

一种新的基于四点插值细分的图像修复算法

本文提出的基于四点插值细分模式的破损图像修复算法和梯度方法相结合,在破损图像修复中得到了较好的结果     

基于混沌的图像置乱加密算法及MATLAB的实现

提出了一种基于混沌映射的图像置乱加密算法。借助MATLAB6.5软件平台编程实现，并研究了加密算法的抗破损能力。实验结果表明：该算法的加密效果良好，图像的抗破损能力强。     

介绍一种打击揉搓式花生脱壳机及其性能影响因素

脱壳是花生食用、加工和种植前的必经工序,同时也是花生加工过程中易造成果仁损伤的主要环节。目前,我国花生脱壳设备还存在适应性差、果仁破伤率高等问题,现有设备不能满足实际需求,果仁破损问题尤为突出,破伤率一般在5～8%,破损的果仁容易脱脂、霉变并易遭到黄曲霉毒素的污染,难以贮藏,通常脱壳机脱出的果仁只能用于榨油或其他     

基于3D技术的货车车体图像检测应用研究

针对货运列车在神华铁路日常的运营过程中,车辆车体受外界及自身磨耗等因素的影响,而造成车体破损、车体仓门缝隙过大,甚至出现严重破损等故障。而故障诊断是实现铁路科学维修决策的重要依据。本文提出了一种利用3D图像检测技术,进行货车车体故障数据处理和诊断的方法。引入主流成像技术,提出一种采用三维图像采集方式,用于覆盖车体整个侧部表明图像的拍摄。进行了车体数据采集、3D深度数据分析、数据测试,并通过实验验证了方法的有效性。     

基于神经网络和不变矩的路面破损图像识别

提出了一种应用BP神经网络识别沥青路面破损图像的图像分割方法.将路面图像等分,用灰度方差值描述子块图像特征,利用BP神经网络对子块图像进行模式分类,并将图像子块模式矩阵的不变矩作为图像的整体特征,在此基础上设计了基于全局优化算法的前馈神经网络分类器,并进行了图像识别试验,对二值图像进行特征提取,提出了学习算法,以加快收敛速度,从而实现图像识别.     

一种基于线性模型拟合的纸币图像边界检测方法

从贝叶斯推理的角度提出一种计算线性模型参数的方法,并将该方法应用到高速纸币清分系统的纸币图像边界检测中,实验结果表明,可以通过少量观测数据实现图像边界线性模型的参数估计,从而完成边界的快速检测。检测精度高,同时针对纸币边缘破损的情况,提出了有效避免旧纸币边缘破损对纸币定位影响的方法。     

两种基于混沌序列的图像加密方法的比较

对基于混沌序列的图像加密算法,以及基于混沌序列的图像置乱加密算法进行了分析、比较.实验结果表明:前者优势在于简单易行,加解密速度快;而后者主要优势在于抗干扰、抗破损性能更佳.     

Elemental classification in multi-detector stem images using image analysis clustering techniques

The multi-detector STEM images form a multi-dimensional measurement space where picture elements of morphologically distinct regions cluster and can be separated for classification using image processing clustering techniques. These images are highly correlated and improve the classification only at the high cost of adding dimensionality. Data reduction techniques which take the class separation into account can be used to compress the useful information carried by these images into a few components to which clustering techniques can be successfully applied. At high magnification, a slight displacement is sometimes observed, which can slightly impair the classification result. The redundancy of the information in the quadrant images suggests their use for phase retrieval - which, added to an energy loss data channel, may greatly improve the classification.     

基于Matlab的岩羊运动图像处理方法的研究

对岩羊运动图像进行数字图像处理,得到比较清晰的岩羊运动图像。文章采用MATLAB处理工具,结合多种数字图像处理方法,并对滤波和直方图均衡的不同方法进行对比和评价,实现数字图像清晰化处理。     

High-resolution wave field reconstruction using a through-focus series of images taken under limited electron dose

The three-dimensional Fourier filtering method and Schiske's Wiener filtering method are compared with the aim of high-resolution wave field reconstruction of an unstained deoxyribonucleic acid (DNA) molecular fiber using a through-focus series of images taken under a limited electron dose. There were some definite differences between the two reconstructed images, although the two kinds of processing are essentially equivalent except for the dimension and the filter used for processing. Through theoretical analyses together with computer simulations, the differences were proved to be primarily due to specimen drift during the experiment. Although the observed structure of the DNA molecular fiber was heavily damaged by electron beam irradiation, reconstructed images by the three-dimensional Fourier filtering method provided higher resolution information on the molecular structure even when relatively large specimen drift was included in the through-focus series. In contrast, in Schiske's Wiener filtering method, the detailed information of the structure was lost because of the drift, although the reconstructed image showed a higher signal-to-noise ratio. The three dimensional Fourier filtering method seems to be more applicable for observing radiation-sensitive materials under an extremely low electron dose, because specimen drift cannot be completely avoided.     

Extraction of sizes and velocities of spray droplets by optical imaging method

In this study, an optical imaging method was developed for the measurements of the sizes and velocities of droplets in sprays. Double-exposure single-frame spray images were captured by the imaging system. An image processing program was developed for the measurements of the sizes and positions of individual particles including separation of the overlapped particles and particle tracking and pairing at two time instants. To recognize and separate overlapping particles, the morphological method based on watershed segmentation as well as separation using the perimeter and convex hull of image was used consecutively. Better results in separation were obtained by utilization of both methods especially for the multiple or heavily-overlapped particles. The match probability method was adopted for particle tracking and pairing after identifying the positions of individual particles and it produced good matching results even for large particles like droplets in sprays. Therefore, the developed optical imaging method could provide a reliable way of analyzing the motion and size distribution of droplets produced by various sprays and atomization devices.     

全自动自准直经纬仪轴系精密伺服驱动系统

针对传统电子经纬仪轴系手动操作的不足,设计轴系直流电机伺服系统以实现其自动化操作.以控制计算机为上位机,负责对加装在电子经纬仪中的面阵CCD相机采集到的图像进行图像处理,得到角度偏差并将此偏差信号输入到加装在电子经纬仪中的dsPIC30 F6012A芯片.由控制芯片依据控制算法运算处理后得到电机驱动信号并经驱动放大后控制电机转动.位置控制算法选用积分分离式PID算法.实验结果表明:加装电机后的电子经纬仪在实现了轴系自动转动的同时能保持其原有测量精度.     

Digital image inpainting and microscopy imaging

A considerable amount of image processing techniques known as inpainting techniques have been recently developed aiming to provide solutions for filling in missing or damaged regions in a digital image. Typical such techniques reconstruct a defined area by using information from its neighborhood, for example, by completing inside the missing region the isophote lines arriving at its boundaries. In this article, we show that inpainting techniques have considerable potential usefulness in microscopy imaging, even though experimenting and using them in this domain has been almost entirely neglected up until now. In this purpose, we experiment the "curvature-preserving" partial differential equations as a solution to inpainting regions in images collected by several optical and scanning probe microscopy techniques. The results achieved are presented along with a discussion on typical problematic scenarios of microscopy imaging for which this type of techniques can provide a viable solution.     

Modified generalized hough transform for 3D image processing with unknown rotation and scaling parameters

An approach to 3D image processing based on the modified generalized Hough transform is proposed. The possibility of processing images with unknown rotation and scaling parameters and images represented by individual fragments of the original image is shown. The efficiency of the proposed algorithms for solving problems of 3D image recognition is studied.     

High resolution X-ray phase contrast synchrotron imaging of normal and ligation damaged rat sciatic nerves

This study was performed to apply synchrotron radiation (SR) imaging to a neuropathologic evaluation technique after treatment of peripheral nerve blocks. A phase contrast synchrotron images of normal and ligation damaged rat sciatic nerve were obtained with an 8 KeV monochromatic beam and 20-mum thick CsI(TI) scintillation crystal. The visual image was magnified using a 20x microscope objective and captured using an analog CCD camera. Obtained images were compared with conventional light microscopic findings from the same nerve samples. By using an edge enhancement effect of phase const with SR, we could easily discriminate each nerve fiber and identify the arrangement of nerve fibers within a whole thickness (about 1 mm in diameter) of peripheral nerve without sectioning and fixation. The composite SR image of a ligation damaged rat sciatic nerve sample showed that the response to nerve injury was different on each side of the site of injury. The SR image of damaged distal lesion showed destruction of neural microarchitecture and typical extensive Wallerian degeneration of nerve fibers as clearly as histologic image. We could get very detailed morphologic data for Wallerian degeneration of nerve fibers by using the SR imaging technique. We believe that the phase contrast synchrotron imaging has great potential as an imaging tool in the bioscience and medical science.     

基于小波与双三次插值的高温锻件红外图像增强方法

提出了一种基于小波变换和双三次插值的图像处理方法,该方法利用离散小波将原图像分解成低频子带与高频子带,然后对原图像及这些子带分别进行双三次插值,同时提出增加一个中间步骤来估算高频子带,即扣除原低分辨率图像和插值后的LL子带中的相同成分而得到不同成分的图像（高频成分）,然后利用不同成分的图像对高频子带LH、HL及HH分别进行校正融合,并通过小波逆变换对这些图像进行融合重构。该方法不仅能够有效地消除图像噪声,改善细节信息,而且能够最大限度地保留边缘信息和增加图像的清晰度。最后运用该方法和传统算法分别对核电主管道高温锻件和小型高温锻件的红外图像进行增强处理,实验结果验证了该方法的可行性及优越性。     

应用数字图像处理技术检测塔式起重机钢丝绳的损伤

介绍1种采用数字图像处理实现塔式起重机钢丝绳损伤检测的新方法,对采集到的钢丝绳图像进行预处理以减少噪声的污染和影响,根据损伤钢丝绳的特征对预处理的图像进行分割,对图像进行数学形态学处理并判断钢丝绳是否存在损伤。