相机位姿估计的加速正交迭代算法

面对需要实时计算的相机位姿估计问题,针对经典的广泛应用的正交迭代算法,提出了一种加速正交迭代算法。其关键思想是将每一次迭代过程规整化,从而提炼出每一次迭代的重复计算,若将此重复计算在迭代开始前提前计算,则可以大幅度的减少迭代过程中的计算量,使得每一次迭代的计算复杂度从O(n)降低为O(1)。因此,可以在更短的时间内迭代更多的次数,从而获得更高的精度。进行了对比实验,结果显示本加速算法计算精度更高,速度更快。并通过实验提出了选择稳健n点透视(RPn P)计算初值,再使用加速正交迭代算法进行迭代运算的方法,在控制点不多的情况下,是一种精度接近最大似然估计,计算速度最快的算法。     

基于机器视觉的移动工件抓取和装配的研究

针对工业生产线上某产品上下盖自动装配的问题,运用运动控制理论与机器视觉相结合,搭建了一套模拟工业生产线产品装配系统。首先利用工业相机对进入视场的工件进行单帧图像采集,获取图像数据;然后利用OpenCV对数字图像进行处理,包括中值滤波、阈值分割、孔洞填充、边缘提取来获取数字图像的特征;之后通过质心法结合矩形拟合计算出工件的质心坐标和偏转角度;最后引导六轴机械臂抓取产品的时候能根据图像处理结果进行纠偏和补偿。该系统采用平面棋盘格标定板对工业相机进行了标定,并且在线进行了验证,实验结果表明,工件抓取和装配成功率为100%,满足工业现场的装配精度要求。     

基于OpenCV的嵌入式自动捡乒乓球系统设计

为了解决在乒乓球比赛和训练中人工手动捡球的问题,从机器视觉的角度出发,综合嵌入式、图像处理和测量控制等技术,基于第三方库OpenCV设计了一种自动捡乒乓球系统。系统是以Contex-A8平台为控制核心,首先通过对云台和USB摄像头的控制获取视频信号;然后通过霍夫圆变换和SVM分类器算法识别出视频中的乒乓球,并根据质心计算出其与系统的相对位置;最后将视频信号中抽象出的位置信息传递给机械运动控制终端,控制小车和机械手将乒乓球捡起,并送到指定位置。实验结果表明本系统能够正确的检测和识别乒乓球,并对其进行定位捡取,效果良好,提高了拾取乒乓球的效率,减轻工作人员的负担。     

Quantitative colorimetric-imaging analysis of nickel in iron meteorites

A quantitative analytical imaging approach for determining the nickel content of metallic meteorites is proposed. The approach uses a digital image of a series of standard solutions of the nickel-dimethylglyoxime coloured chelate and a meteorite sample solution subjected to the same treatment as the nickel standards for quantitation. The image is processed with suitable software to assign a colour-dependent numerical value (analytical signal) to each standard. Such a value is directly proportional to the analyte concentration, which facilitates construction of a calibration graph where the value for the unknown sample can be interpolated to calculate the nickel content of the meteorite. The results thus obtained were validated by comparison with the official, ISO-endorsed spectrophotometric method for nickel.The proposed method is fairly simple and inexpensive; in fact, it uses a commercially available digital camera as measuring instrument and the images it provides are processed with highly user-friendly public domain software (specifically, ImageJ, developed by the National Institutes of Health and freely available for download on the Internet).In a scenario dominated by increasingly sophisticated and expensive equipment, the proposed method provides a cost-effective alternative based on simple, robust hardware that is affordable and can be readily accessed worldwide. This can be especially advantageous for countries were available resources for analytical equipment investments are scant. The proposed method is essentially an adaptation of classical chemical analysis to current, straightforward, robust, cost-effective instrumentation.     

Environmental monitoring using a conventional photographic digital camera for multianalyte disposable optical sensors

The primary interest of this study concerns the use of an inexpensive photographic digital camera as the detection system, using its own flash as the source of light to present a new analytical procedure to measure disposable multianalyte optical sensors for potassium, magnesium, hardness and conventional pH test strips. The camera arrangement was designed in a fixed position over an optical board with controllable ambient conditions. After acquiring the digital image, the analytical information contained in each test zone is analyzed using theRGB colour space. Reflectance measurements were developed to study the colourimetric and spectral characteristics of the test zones. We obtained the following application ranges and precision in terms of relative standard deviation (RSD %): for potassium from 3.2 × 10⁻⁷ to 0.1 M with a precision between 3.3 and 4.0%, for magnesium from 2.7 × 10⁻⁶ to 1.5 M showing a precision between 4.7 and 7.8% and finally for hardness from 4.3 × 10⁻² to 200,000 mg L⁻¹ CaCO₃ and between 5.1 and 7.0%. Moreover, the analytical characteristics of several optical procedures were compared with the results presented here. The proposed method was statistically validated against a reference procedure using samples of water from different sources and beverages, indicating that there are no significant statistical differences at a 95% confidence level.     

Visualization and analysis of wheel camber angle effect for slope traversability using an in-wheel camera

This paper visualizes and analyzes an effect of a wheel camber angle for the slope traversability in sandy terrain. An in-wheel camera developed in this work captures the wheel-soil contact phenomenon generated beneath the wheel through a transparent section of the wheel surface. The images taken by the camera are then analyzed using the particle image velocimetry. The soil flows with various wheel camber angles are analyzed with regard to the soil failure observed on the slope surface. The analysis reveals that the slope failure and soil accumulation in front of the wheel significantly affect the wheel forces and distributions of the wheel sinkage in the wheel width direction. Further, the side force of the wheel in traversing a slope decreases as the slip ratio increases because the shear stress in the slope downward direction decreases owing to the slope failure.     

基于数码相片和颜色空间转换的滨海土壤盐渍化定量估算

土壤盐渍化是土壤退化的重要原因之一,快速精确地监测土壤盐渍化对农业可持续发展和生态环境保护有积极作用。提出一种基于数码相片的滨海地区表层土壤盐分的定量估算方法,旨在复杂天气状况下快捷方便的获取土壤盐分信息。以江苏盐城沿海地区裸露地表土壤作为研究对象,在晴天和多云天气下全天时采样和拍照获取52个土壤样本和相片。土壤样品通过室内测试获取土壤电导率（EC）,pH值和土壤含水量等参数。利用RStudio软件对土壤相片进行处理,首先从相片中提取RGB三种颜色参数,再通过颜色空间转换关系计算另外5种颜色空间（HIS,CIEXYZ,CIELAB,CIELUV和CIELCH）,每个颜色空间有三个颜色参数,加上RGB颜色空间共有18个颜色参数,其中CIELAB,CIELUV和CIELCH中L参量表示的意义和数值相同,因此6个颜色空间共有16个颜色参数。土壤电导率与颜色参数相关分析结果表明,相片颜色的纯度和亮度与土壤电导率之间的相关系数较高,并达到了极显著水平,相片颜色的色相与土壤电导率之间的相关性较低,且未达到显著水平。随机抽取70%的样本数据并用随机森林方法对土壤盐分含量进行建模,采用留一法（LOOCV）进行交叉验证,再用余下30%的样本数据进行精度检验,重复100次以获取精度最高的模型。最终获取估算土壤盐分的随机森林模型,验证集数据的模型精度达到R2_val=0.75,RMSE_val=3.52,RPD_val=2.02。对颜色参数进行重要性分析发现,颜色纯度对模型的重要性最大,其次是颜色亮度,色相的贡献较小。综上,利用数码相机获取表层土壤相片,通过颜色空间转换得到的颜色参数为有效估算滨海土壤盐分含量提供一个新思路。该研究对近地表参数定量估算提供了新视角,将来结合无人机平台能够为精准农业和滨海生态环境的精准管理提供技术支持和有效手段。     

TDICCD拼接相机的像元响应非均匀性校正方法

为了解决光学TDICCD成像系统在拼接模式下的像元响应非均匀性问题,研究了TDICCD像元校正的原理和实现方法。提出了视频处理器在不同增益、不同偏置和TDICCD在不同积分时间下进行像元校正的方法,设计了在现场可编程门阵列（FPGA）平台下进行了程序实现和验证的方法。实验数据分析表明：成像系统单片TDICCD的非均匀性由4.72%降低到0.27%,恶劣环境下TDICCD成像系统的图像非均匀性可以降低2.55%。该像元级校正算法简单,可靠性高,能够满足星上成像的要求;在不同的增益、偏置和积分时间下能够很好地解决TDICCD成像系统的像元响应非均匀性问题。     

电视摄象机测量误差探讨

本文简述了用电视摄象机进行工业尺寸测量的微分法与拟合法的基本原理,同时详细的讨论了各自所能达到的精度极限。本文用实验验证了理论计算的结果,最后对二维工件测量时,如何根据工件的不同形状与位置以及光学系统的特点,提出了在测量时工件的放置应该注意的问题以及对工件拐角处如何进行处理的方法。     

多通道CCD摄像机的数据采集和图像显示

针对多通道CCD摄像机 ,提出了一种多线程数据实时存储的方法。利用多线程提高硬盘读写速度 ,实现高速、大容量CCD摄像机数据的硬盘实时存储 ,并利用定时中断将多通道信号定时分块存储入硬盘 ,存储的CCD摄像机数据被转换为图像文件加以显示。系统的CCD摄像机通道达到 5个 ,每个通道的数据率达到 10MBit/s。通过实验 ,验证了该方法的正确性和有效性