重叠苹果果实的分离识别方法

针对苹果采摘机器人重叠果实识别误差较大的问题,设计了一种分离识别方法。首先在苹果图像分割获取其二值果实区域的基础上,基于横、纵投影图实现重叠形态果实的判别,而后基于边缘曲线通过SUSAN算法检测果实轮廓上的角点,再通过迭代腐蚀和瓶颈准则挑选重叠果实的分离点,并采用Bresenham算法连接分离点实现重叠果实的分离。提取分离果实边缘曲线的有效轮廓后,通过改进的随机Hough算法拟合果实圆心及半径。最后选择15幅重叠果实区域二值图像,通过不同角点检测计算结果的比较,验证了SUSAN算法相比于其他角点检测方法更为有效;通过改进的随机Hough算法识别11幅图像中的21个果实,其圆心相对误差平均值、半径相对误差平均值和相对偏差平均值分别为6.90%、4.12%和6.07%,比传统Hough算法分别低4.03、2.75、1.14个百分点,说明改进的随机Hough算法得到的拟合圆更接近实际苹果果实区域。     

3类朝天椒重金属健康风险评价与镉吸收累积差异

为探究遵义市朝天椒的健康风险及其吸收累积风险元素镉(Cd)的差异，采用目标危害系数法评价了3类朝天椒(按形状分为指型、锥型与圆型)果实重金属镉(Cd)、砷(As)、铅(Pb)、铜(Cu)、锌(Zn)的健康风险，并通过盆栽试验探究3类朝天椒对Cd的吸收累积差异的机制。结果表明，研究区3类朝天椒仅存在Cd超标现象，指型朝天椒、锥型朝天椒与圆型朝天椒的果实对土壤Cd的富集系数(biological enrichment factor,BCF)分别为0.35、0.25、0.15。食用圆型朝天椒危害指数最低，食用指型朝天椒危害指数最高，健康风险主要来源于Cd。盆栽试验的3类朝天椒果实Cd含量表现为：指型朝天椒>锥型朝天椒>圆型朝天椒，前两者果实Cd含量超标(>0.1 mg·kg^-1)，而圆型朝天椒果实Cd含量未超标；圆型朝天椒对土壤Cd的富集系数及其茎向果实的Cd转运系数(transport factor, TF) TF_果/茎和叶向果实的Cd转运系数(TF_果/叶)均显著低于指型朝天椒。指型朝天椒与圆型朝天椒的根向...     

基于图像的番茄识别与几何尺寸测量

介绍了一种基于图像的番茄识别算法,比较了几种常用的边缘提取算法的效果,最终选择Canny算子进行边缘提取,在对番茄进行轮廓提取的基础上,以像素为单位,对番茄的面积、形心、半径等参数进行统计,然后用一个圆来完成对番茄的拟合。这种识别算法可以识别各种光照条件下的果实,并且可以在一定程度上恢复番茄被枝叶遮挡住的部分。最后在Matlab上进行了算法设计和几何尺寸测量。     

基于开源项目的苹果采摘系统设计及视觉算法改进

以机器人操作系统(robot operating system,ROS)为框架,以开源项目树莓派微型硬件作为图像处理模块,Arduino作为机械臂运动控制模块;基于自适应参数输入的霍夫圆检测算法,实现一种USB摄像头的双目视觉苹果采摘系统,保证算法实时性高的前提下识别率较高。以轮廓完整、小色块去除为优化策略,最大、最小圆面积自动估算,实现对霍夫圆检测输入参数的自适应控制。对于采摘顺序规划,通过竞争策略,每次筛选出最优果实,从而实现多目标的采摘顺序确定。结果表明,基于本策略的自适应霍夫圆检测试验算法对遮挡状况下的小规模成簇果实具有较高识别率。此外,借助众灵科技有限公司的开源ROS五自由度机械臂通过简易抓取试验完成了算法可实施性验证。     

基于量子粒子群的柑橘果实图像约束聚类分割算法

为降低果蔬采摘机器人图像处理系统的分割错误率,提出基于量子粒子群的柑橘果实图像约束聚类分割算法。首先,将读入的RGB模式柑橘果实彩色图像转换成灰度图像;然后采用具有随机遍历性的量子粒子群全局寻优策略,结合对图像中的少量模糊和不确定点具有良好适应性的约束聚类方法,对柑橘图像进行分割。从采集的柑橘果实彩色图像中分别取遮挡柑橘图像与重叠柑橘图像,使用本算法进行分割试验,并分别与OTSU算法、模糊聚类图像分割算法、量子粒子群图像分割算法进行比较试验。结果表明,本算法具有更低的分割错误率以及更高的峰值信噪比。     

黑贝1号圆茄

黑贝1号圆茄，母本为二苠茄高代自交纯化分离的品质98—24，父本为98—8品系，2001-2002年春季进行区域试验与示范推广，同时对黑贝1号圆茄的异常果、耐寒性、果实褐变情况进行了研究。     

露地番茄病烂果的防治对策

(接上期) 三、真菌性病烂果的识别与防治 这类病烂果由病原真菌侵染所致,可以传染,受害果实上出现不同形状的病斑和霉层.根据果实的受害期,可分为两类,第1类主要为害成熟果实,如炭疽病、根霉果腐病、酸腐病等,第2类主要为害青果,如绵腐病,绵疫病、晚疫病、圆纹病、灰斑病、黑斑病、早疫病等.     

自然环境下苹果作业机器人双果重叠目标侦测方法

为进一步提升苹果的侦测精度,从而提高苹果作业机器人的果实作业效率,以果园重叠苹果为对象,研究了自然环境下双果重叠目标的机器侦测方法。首先将采集到的苹果图像在Lab色彩空间中利用K-means聚类算法提取重叠苹果目标区域;其次在得到重叠苹果边界上的Harris角点后,通过关键角点检测算法定位苹果重叠部分的果实轮廓所在区域,并利用Canny边缘检测算子提取出苹果重叠部分的果实轮廓;然后利用Y型节点搜索算法实现重叠苹果目标的单果轮廓分离,并得到未遮挡果实的完整轮廓;最终利用距离最小二乘算法对被遮挡苹果目标进行果实轮廓重建。为验证方法的有效性,将Hough变换法、Spline样条内插法的重叠果实侦测结果与本文方法所得结果进行对比。试验结果显示,本文方法不仅可以侦测出未遮挡果实的完整轮廓,同时对被遮挡苹果目标也有较好的轮廓重建效果,其果实平均重合度和定位误差分别为95.43%和4.44%,侦测性能明显优于前两种方法,表明本文方法可以较好实现自然环境下苹果双果重叠目标的侦测,该结果为苹果作业机器人多果重叠目标的自动化侦测提供参考。     

霞光油桃×NF油蟠桃组合F1代果实主要性状的遗传倾向

以油桃品种霞光和油蟠桃品系NF杂交组合为试材,分析了F1代果实主要性状遗传倾向.进一步证实了果形遗传符合1对等位基因的分离规律,蟠桃对圆桃为完全显性;果实成熟期属数量性状,后代出现较广泛的分离,45.9%的单株成熟期早于双亲.由于在有性过程中非加性效应的解体,果实大小低于双亲,但仍有9.4%的油蟠桃单株超过油蟠桃亲本.杂种后代果实风味呈现偏甜的趋势,可溶性固形物含量平均值低于双亲,超过双亲的油蟠桃单株仅占4.1%.果重与果形之间呈极显著负相关,可溶性固形物含量与果实成熟期、风味呈极显著正相关,pH值与风味呈极显著正相关.     

基于外部特征信息的番茄果实质量预测模型

为实现番茄果实质量的预测,提出一种应用计算机视觉技术对番茄果实质量进行自动识别的方法。采用Matlab平台构建算法,利用数学形态运算及图像局部性质运算等构成的识别算法,对番茄样本图像进行预处理,提取番茄图像投影面积、轮廓周长、最大内接圆直径和最小外接圆直径外部特征参数,分别建立番茄果实质量与4个特征参数之间的一元线性、二阶多项式、幂指和多元线性预测模型。试验结果显示：采用多元线性预测模型预测结果最佳,其决定系数为0.926 7, 标准差为4.32; 利用检验样本对预测模型进行验证,预测果实质量与实际果实质量的绝对误差均值为3.260 g, 相对误差均值为1.535%。 研究结果表明,基于计算机视觉技术外部特征信息的番茄果实质量预测方法是可行的。     

Polar wandering on Mars?

Polar wandering during the past 10(8) years may be recorded by unique quasi-circular structures in the polar regions of Mars. Polar wandering on Mars is likely if deep convection is involved in the origin of the very large constructional volcanic features located near the equator. The magnitude of the nonhydrostatic low order components of the gravity field and their correlation with the equatorial volcanic features may be additional evidence of deep convection and associated polar wandering.     

Radar soundings of the subsurface of Mars

The martian subsurface has been probed to kilometer depths by the Mars Advanced Radar for Subsurface and Ionospheric Sounding instrument aboard the Mars Express orbiter. Signals penetrate the polar layered deposits, probably imaging the base of the deposits. Data from the northern lowlands of Chryse Planitia have revealed a shallowly buried quasi-circular structure about 250 kilometers in diameter that is interpreted to be an impact basin. In addition, a planar reflector associated with the basin structure may indicate the presence of a low-loss deposit that is more than 1 kilometer thick.     

一种粘连谷物图像分割及杂质识别算法开发

本文提出了一种自动分割粘连谷物并识别杂质的算法。该算法首先使用基于形态学多尺度分解的分水岭算法对粘连颗粒进行分割,接着提取各个颗粒的形态特征和颜色特征,然后计算上述样本颗粒的特征值与完好谷物的特征值之间的马氏距离,并比较它们的马氏距离与设定的阈值,来识别混杂在谷物中的杂质。通过对5种谷粒(普通大米、粗米、糙米、普通大麦、糯麦)的实验,结果表明该算法取得了较好的分割与识别效果,为谷物质量分级的评定提供了一种快速有效地检测谷物产品杂质率的方法。     

两个蓝莓品种离体叶片不定芽再生体系的建立

以Sunrise和Geo两个蓝莓品种试管苗叶片为外植体材料,以改良的WPM为基本培养基,研究了暗培养时间、叶片放置方式和不同浓度激素组合（ZT、IBA、TDZ）对2个蓝莓品种叶片不定芽再生的影响。结果表明：最适的暗培养时间为14d,近轴面放置效果最好,不同基因型不定芽再生能力不同,Sunrise品种在WPM＋4．0mg／L ZT＋0．3mg／L IBA＋1．0mg／L TDZ＋20g／L蔗糖＋6g／L琼脂培养基中生长最好,离体叶片不定芽再生频率达100％,再生不定芽数可达4．5个以上;Geo品种最适宜培养基为WPM＋4．0mg／LZT＋0．3mg／LIBA＋20g/L蔗糖＋6g／L琼脂,叶片不定芽再生频率达到100％。     

木材表面粗糙度的分析

该文采用探针法测定水曲柳、毛白杨和杉木不同加工表面的粗糙度,分析了影响木材表面粗糙度的因素. 并采用心理分析方法讨论了木材表面的粗糙感,以及触觉心理量与视觉心理量的关系. 为获得不同树种、切削加工方式、切面对木材工件表面粗糙度的影响,该实验分别选择了锯切、铣削、刨切和磨削加工的木材工件表面进行表面的粗糙度和粗糙感触觉心理量、视觉心理量的测量. 研究结果表明,树种、切削方式、切面和木材的表面组织构造会影响木材的表面粗糙度;表面粗糙度值因为加工方法不同而不同,随着加工精度的提高而降低. 水曲柳、毛白杨和杉木表面粗糙感的触觉心理量与视觉心理量呈正相关,针叶材表面粗糙感的触觉心理量与视觉心理量的相关性比阔叶材的大.       

基于脉冲耦合神经网络的粘连玉米种子图像分割

为解决玉米种子内部机械裂纹检测过程中存在的种子间粘连问题,提出一种基于自适应脉冲耦合神经网络(Pulse coupled neural net)模型与熵值最大原则相结合的图像分割算法。运用直方图均衡化和布特沃斯低通滤波器进行频域增强预处理,以提高玉米种子与图像背景的对比度;运用PCNN模型,结合最大熵值原则对预处理后的粘连玉米种子图像进行分割,并引入图像像素的拉普拉斯能量(Energy of laplace)作为PCNN网络各神经元之间的连接系数,以增强图像分割效果;采用维纳滤波和数学形态学对分割后存在的噪声和断点进行处理,得到最终的分割效果。试验结果表明:PCNN与熵值最大原则相结合的图像分割算法的分割准确率为92.5%,运行时间为1.166 2s,分割准确率高于改进分水岭算法、OTSU算法和最大熵直方图分割算法,用时略长于其他分割算法。     

Large quasi-circular features beneath frost on triton

Specially processed Voyager 2 images of Neptune's largest moon, Triton, reveal three large quasi-circular features ranging in diameter from 280 to 935 kilometers within Triton's equatorial region. The largest of these features contains a central, irregularly shaped area of comparatively low albedo about 380 kilometers in diameter, surrounded by crudely concentric annuli of higher albedo materials. None of the features exhibit significant topographic expression, and all appear to be primarily albedo markings. The features are located within a broad equatorial band of anomalously transparent frost that renders them nearly invisible at the large phase angles (alpha > 90 degrees ) at which Voyager obtained its highest resolution coverage of Triton. The features can be discerned at smaller phase angles (alpha = 66 degrees ) at which the frost only partially masks underlying albedo contrasts. The origin of the features is uncertain but may have involved regional cryovolcanic activity.     

Styles of volcanism on venus: new arecibo high resolution radar data

Arecibo high-resolution (1.5 to 2 km) radar data of Venus for the area extending from Beta Regio to western Eisila Regio provide strong evidence that the mountains in Beta and Eisila Regiones and plains in and adjacent to Guinevere Planitia are of volcanic origin. Recognized styles of volcanism include large volcanic edifices on the Beta and Eisila rises related to regional structural trends, plains with multiple source vents and a mottled appearance due to the ponding of volcanic flows, and plains with bright features surrounded by extensive quasi-circular radar-dark halos. The high density of volcanic vents in the plains suggests that heat loss by abundant and widely distributed plains volcanism may be more significant than previously recognized. The low density of impact craters greater than 15 km in diameter in this region compared to the average density for the higher northern latitudes suggests that the plains have a younger age.     

Support vector machine approach to real-time inspection of biscuits on moving conveyor belt

An intelligent system for colour inspection of biscuit products is proposed. In this system, the state-of-the-art classification techniques based on Support Vector Machines (SVM) and Wilk's λ analysis were used to classify biscuits into one of four distinct groups: under-baked, moderately baked, over-baked, and substantially over-baked. The accuracy of the system was compared with standard discriminant analysis using both direct and multi-step classifications. It was discovered that the radial basis SVM after Wilk's λ was more precise in classification compared to other classifiers. Real-time implementation was achieved by means of multi-core processor with advanced multiple-buffering and multithreading algorithms. The system resulted in correct classification rate of more than 96% for stationary and moving biscuits at 9 m/min. It was discovered that touching and non-touching biscuits did not significantly interfere with accurate assessment of baking. However, image processing of touching biscuits was considerably slower compared to non-touching biscuits, averaging at 36.3 ms and 9.0 ms, respectively. The decrease in speed was due to the complexity of the watershed-based algorithm used to segment touching biscuits. This image computing platform can potentially support the requirements of the high-volume biscuit production.