基于多结构鲁棒估计的虹膜外边缘定位方法

在Adaboost检测人脸区域的基础上，提出了基于多结构鲁棒估计的虹膜外边缘定位方法．通过对200幅单人脸的图像数据集的仿真实验，结果表明，本方法相比于传统的RANSAC模型生成算法以及Hough变换等方法，能得到更精确的定位结果，而且能加快有效的模型生成，虹膜外边缘定位精度达到94％．     

浅谈沥青路面的病害与防治

本文作者对沥青路面的病害成因作了一般性分析及论述,并提出了几项防治措施。     

图像分割技术在木材表面缺陷识别中的应用

阐述了图像分割技术在图像处理及分析中重要意义,以及主要的图像分割技术。以木材表面缺陷为主要研究对象,利用微分算子边缘检测、最优迭代阈值分割及形态学方法针对具有代表性缺陷死节和虫眼进行分割处理。     

Avoiding edge effects in agroforestry experiments; the use of neighbour-balanced designs and guard areas

A number of different types of edge effects on experimental plots are defined, and it is shown that they are frequently important in agroforestry experiments. Methods of eliminating edge effects, either at the design or the analysis stage, are described. Computer simulations, based on uniformity data, are used to investigate the efficacy of neighbour-balanced designs, with and without neighbour terms, for reducing the problem of neighbour effects between different treatments. It is concluded that, while neighbour-balanced designs may be useful in certain situations or in combination with other methods, it is preferable to eliminate edge effects by using adequate guard rows. Recommendations are therefore made concerning the correct use of guard areas in agroforestry experimentation.     

三维机械CAD系统中的参数化特征库技术

介绍了在三维机械CAD系统Solid Edge中，建立常用零件参数化特征库的技术以及参数化特征的使用方法，着重探讨了将驱动尺寸与变量表结合起来提高常用零件建模效率的技术。     

青石山区堰边核桃良种丰产栽培技术研究

利用青石山区梯田堰边栽植1～2a生核桃实生苗，株距4～5m，行距不等，每公顷平均375株。缓苗后进行良种嫁接。主要品种有元丰、香铃、鲁光、阿九等嫁接第2年结果，第3年有一定产量。每666．7m^2核桃产量3a生为33．5k，5a生为915kg，9a生为218．25kg，12a生为273．25kg。堰边良种核桃的综合经济效益分别为对照的1．5～4．9倍阐述了主要优质丰产栽培技术，包括实生苗建园与幼树嫁接良种、合理整形修剪、肥水管理与合理间作、适期采收、坚果处理及病虫害防治等。     

海南岛尖峰岭次生热带山地雨林的边缘效应

对海南岛尖峰岭热带山地雨林采伐迹地次生林及周围的保留林进行了群落调查．组建了一个基于分割线段模型的边缘效应测度公式：Y=α e0[1-(D／Dmax)](式中Y为与边缘垂直距离D处的测定指标值，e0为边缘效应强度的最大值，Dmax为边缘效应作用的最大距离，α为常数)，研究了热带山地雨林群落的边缘效应，结果表明：(1)用组建的测度公式来分析热带山地雨林采伐迹地边缘的次生林及保留林的边缘效应强度及其作用距离，结果是有效和可靠的；(2)热带山地雨林采伐后形成的边缘，其效应的作用距离一般都不超过15m；(3)边缘对保留林和次生林的效应没有显著差异。     

Patterns of fish movement at a desert river confluence

Quantifying fish movements in river networks helps identify critical habitat needs and how they change with environmental conditions. Some of the challenges in tracking fish movements can be overcome with the use of passive integrated transponder (PIT) tagging and antennas. We used PIT technology to test predictions of movement behaviour for four fish species at a mainstem–tributary confluence zone in an arid‐land river system. Specifically, we focused on the McElmo Creek tributary confluence with the San Juan River in south‐western Utah, USA. We quantified variation in species occurrences at this confluence zone from May 2012 to December 2015 relative to temporal and environmental conditions. We considered occurrences among species relative to tagging origins (tributary versus mainstem), season and time of day. Generally, fishes tagged in the focal tributary were more likely to be detected compared to fish tagged in the mainstem river or other tributaries. Additionally, adults were most likely to be detected across multiple years compared to subadults. Based on a Random Forests model, the best performing environmental variables for predicting seasonal detections included mainstem discharge during run‐off season (razorback sucker Xyrauchen texanus), tributary discharge during monsoon season (Colorado pikeminnow Ptychocheilus lucius) and mainstem water temperature (flannelmouth sucker Catostomus latipinnis and channel catfish Ictalurus punctatus). The variable responses by endemic and introduced fishes indicate tributary habitats provide several key functions within a fish community including spawning, rearing, foraging and refuge.     

基于边缘检测和区域定位的玉米根茎导航线提取方法

基于玉米根茎图像信息，提出一种基于边缘检测和区域定位的玉米根茎导航线提取方法。首先，利用最大类间方差法自动分割2G-R-B灰度图像，得到二值化图像，结合形态学处理、位置/面积去噪方法提高二值化图像质量，对去噪图像按列累加得到列像素累加曲线；针对传统方法得到的特征点中伪特征点较多的问题，引入高斯滤波器平滑累加曲线，并运用极值法减少玉米根茎伪特征点的干扰，在提取玉米茎秆边缘直线时，提出基于最远茎秆成像宽度的双侧边缘判别思路，通过扫描每条边缘直线的四边形封闭邻域有效剔除伪边缘直线；最后，根据边缘直线二次定位玉米的根茎区域范围，并剔除伪特征点，采用最小二乘线性拟合方法准确提取导航线。试验表明，本文算法处理一幅1280像素×720像素图像耗时约236ms，特征点拟合准确率为92%。与传统方法相比，本文算法精度高、实时性好，在缺苗、杂草较多和株距不标准的情况下仍具有较强的鲁棒性。