Simultaneous HPLC quantification of total cholesterol,tocopherols and β-carotene in Barrosã-PDO veal

A simple, rapid and sensitive procedure for the simultaneous determination of total cholesterol, tocopherols and β-carotene in meat is described. The method involves a direct saponification of the meat, a single n-hexane extraction and the analysis of the extracted compounds by normal-phase HPLC, using fluorescence (tocopherols) and UV–Vis photodiode array (cholesterol and β-carotene) detections in tandem. Rates of recovery of spiked meat samples were 93% for cholesterol, 83–86% for (α-, β- and -γ) tocopherols and 89% for β-carotene. Repeatabilities were high (CV < 6%) for all determined compounds, except for δ-tocopherol. This tocopherol, which is not usually present in meat, showed a much lower recovery percentage (73%) and repeatability (12.8%). This methodology was applied for the quantification of total cholesterol, tocopherols and β-carotene in three muscles (longissimus thoracis, longissimus lumborum and semitendinosus) of the Portuguese traditional Barrosã-PDO veal, obtained from autochthonous calves fed extensively during summer (with the least abundant green pastures) and slaughtered in early autumn (October). Barrosã-PDO veal showed median contents of total cholesterol (0.50–0.56 mg/g) and, depending on the analysed muscle, moderate to high contents of α-tocopherol (3.3–3.9 μg/g) and β-carotene (0.07–0.09 μg/g), suggesting an high sensorial and hygienic quality.     

面向旋翼无人机的高压输电线在线检测方法

面向旋翼无人机电力巡线的高压输电线在线检测对算法实时性、鲁棒性要求高的问题,提出了一种基于边界搜索和Radon变换(BSRT)的高压输电线识别算法.分析了高压输电线在图像中的边界贯穿特征,提出了以图像四条边界作为起始点的搜索策略,给出了边界约束下的直线特征Radon变换能量函数和求解方法.复杂度分析结果表明本算法与经典的Radon算法相比,复杂度降低了一个数量级.以人工合成图像和无人机实际航拍图像,对本算法、Radon算法和LSD算法在实时性和有效性方面进行了实验对比分析,实验结果表明,本算法的处理速度较Radon算法有很大的提高,与LSD算法的处理速度基本处于同一量级,但本算法的高压输电线检测精度大幅优于传统的Radon和LSD算法.理论分析及实验结果证明,提出的BSRT方法有效地解决了经典Radon算法的高复杂度和LSD算法的复杂背景高敏感性的问题,具有较好的应用价值.     

基于机器视觉的转辙机缺口检测方法研究

针对现有转辙机缺口检测方法所存在的问题,本文介绍了一种基于机器视觉技术的转辙机缺口检测基本方法及其原理和步骤。文章详细论述了LSD算法的工作原理,同时本文作者对该算法做出了改进。最后将改进的LSD算法应用在转辙机缺口检测当中,结果表明该检测算法具有良好的稳定性和快速性,并且能够很好的解决转辙机缺口检测的问题。     

基于深度学习增强的LSD杆塔倾斜度检测

杆塔的倾倒会对整个电网产生严重的破坏并威胁周围居民的生命安全,电力巡检无人机利用计算机视觉对杆塔进行巡检既节省了人力资源又显著地提高了电网的巡检效率.为了国网巡检人员在杆塔倾倒前及时得到预警,对电力巡检无人机中的基于计算机视觉的杆塔倾斜检测算法进行了研究,设计了一种基于YOLOv3的深度神经网络结合LSD线段提取方法对杆塔的倾斜进行检测.利用在山西电网无人机实际巡检的杆塔图片制作杆塔的VOC2007数据集并利用YOLOv3神经网络对杆塔进行目标检测,并将检测后得到的Bounding box根据网络训练后的mIOU参数进行微调后作为LSD检测的ROI.接着,该方法在ROI中将检测的线段进行过滤和融合,根据杆塔特点进行杆塔的二次识别.最后利用得到的杆塔外边线做出该方向上的杆塔中线并计算杆塔在该方向的倾斜度.该文中实验利用山西国网电力公司提供的数据进行验证,结果表明,杆塔的倾斜检测效果在各种拍摄高度和背景干扰下都较为精确,杆塔目标的正确识别率达到97％,倾斜度检测平均误差小于0.85°.     

虚拟膝关节镜手术中保持单元质量的体网格形变方法

在虚拟膝关节手术中,需要对膝关节进行大范围形变的实时模拟。本文针对四面体网格的膝关节模型,提出了采用LSD度量建立形变能量,然后将带约束的最优化问题转化为不带约束的最优化问题,最后通过带Armijo线性查找的非精确牛顿法求解最优化问题。在求解过程中,通过预估未知点的位置,减少迭代步数,提高了算法的效率。这种方法具有较好的保体积性,同时保证形变后的四面体网格不出现体元翻转和退化。该方法也能推广应用于其它类似的关节弯曲运动的变形中。     

黄土丘陵沟壑区退化生态系统恢复的研究

在对黄土丘陵沟壑区现有主要退化生态系统类型进行科学分类的基础上,以时间序列和空间序列相结合,进行植物群落动态变化及其演替规律的观察和研究,结果表明:退耕地植物群落演替趋势是退耕地→草本入侵→灌木入侵→乔木入侵→乔木群落.     

基于点云技术的景观空间密度量化方法研究

点云技术的发展为景观空间的精确量化与精准设计提供了重要机遇。选取过往研究中多通过人工实地调研与经验估算的“景观空间密度”指标,利用点云等数字技术优化其量化方法。首先,以无人机倾斜摄影与地面激光雷达扫描生成的点云模型为基础,通过点云体元化得到景观要素的三维形态,进而计算出总体空间密度;其次,利用Grasshopper对空间密度的三维分布加以可视化;最后以东南大学梅庵周边景观空间为例,探讨该方法的实用性。由此发现,基于点云技术的景观空间密度量化方法在数据采集处理的效率、计算结果的精度以及结果的可读性等方面有显著优势,有利于更加直观、准确地为景观空间设计提供针对性的优化策略。     

基于直线检测的棋盘格角点自动提取

针对现有棋盘格角点自动提取算法鲁棒性不足的缺点,提出了使用直线检测算法(LSD)来进行角点的自动提取。首先,使用LSD算法对标定图像进行处理,得出包含棋盘格子边缘的所有直线,并分别对所得直线的长度和角度进行伪排序以去除伪格子边缘。然后,对剩余直线的边缘端点进行近邻合并,得到角点的初始坐标并进行亚像素优化。最后,使用能量法对角点进行棋盘格结构复原与排序。实验结果显示:该方法可以正确提取含有阴影和噪声的玻璃材质标定板图像中的角点。与改进的Harris方法的角点坐标提取精度对比试验得到其最大偏差小于0.2pixel,平均偏差小于0.15pixel,表明该方法具有较高的鲁棒性且定位精度与改进的Harris方法相当,可用于工程实际中环境光源变化较大的场合。     

LSD斜井滑模施工技术

介绍了一种新型的斜井混凝土衬砌施工工艺——LSD斜井滑模系统。该系统是对传统的斜井混凝土衬砌滑模施工工艺的一种革新,在保证施工质量的基础上提高了施工效率,降低了施工成本,获得了显著的经济效益。该系统属国内外首创,已获得国家发明专利,并已在一些大中型水利水电工程施工中得到成功应用。     

Effect of various packaging materials on storage stability of refined,bleached, deodorized palm oil

Refined, bleached, deodorized palm oil (RBD palm oil) was packaged in lacquered metal cans (LMC), green glass bottles (GGB), amber glass bottles (AGB), clear glass bottles (CGB), clear plastic bottles (CPB) and sealed polyethylene film (POLET), and stored in direct sunlight (40 ± 1°C) or in darkness (27 ± 1°C). Measurements of free fatty acid (FFA), peroxide value and anisidine value, at 14-day intervals for a period of 98 days, were used to assess the stability of the oil towards hydrolytic and oxidative deterioration. Total oxidation values for packaged oils stored in direct sunlight showed that LMC gave the greatest protection against oxidative deterioration, followed by GGB and AGB. POLET offered the least protection to the oil against oxidative deterioration, while CPB and CGB proved superior to POLET but inferior to GGB and AGB. For storage in the dark, LMC, AGB and GGB gave the greatest protection to RBD palm oil against oxidative deterioration, with no significant statistical differences between them, while CPB, CGB and POLET followed in that order, with significant differences between their respective abilities to protect the oil against oxidative deterioration. Oils packaged in CPB gave the highest FFA levels (statistically significant). The investigations clearly indicated that LMC is superior to all other tested packaging materials in offering maximum protection to RBD palm oil against hydrolytic and oxidative deterioration. Amber and green glass bottles could serve as viable alternatives to LMC, while CGB and CPB could be tolerated as suitable packaging systems for RBD palm oil. However, the study also clearly showed that POLET is unsatisfactory for use as packaging material for RBD palm oil.