基于KNN方法的大兴安岭地区森林地上碳储量遥感估算

【目的】采用 KNN方法进行碳储量估测,并对估测后的数据进行各种校正处理,绘制森林地上碳储量的空间分布图,为我国森林碳储量和固碳潜力的研究提供基础数据和科学依据。【方法】以黑龙江省大兴安岭为研究区(50°05'—53°33'N,121°11'—127°01'E),基于2010年森林资源连续清查固定样地和同年 Landsat5 TM 影像数据,利用 k－邻近法( KNN)在像素级水平上对森林地上碳储量进行估算。采用多准则方法分东、南、北和中4个区域对样地坐标和其对应的影像光谱值进行坐标重配准,并根据实测样地数据对坐标重配置前后不同林分类型地上碳储量估测精度进行评价;针对 KNN方法像素级估测结果存在明显的高值区域低估和低值区域高估现象,应用直方图匹配方法对估测结果进行变动范围调整;并根据样地实测碳储量和 KNN 估测值间的回归关系对调整后的结果分区域进行进一步匹配校正后处理,绘制森林碳储量的空间分布图。【结果】总体来说,本研究区域像元尺度KNN估测的欧式距离优于马氏距离,均方根误差随着最邻近值 k的增大而降低,当 k大于6时变化缓慢,并逐渐趋于稳定;坐标误差校正后,各林分类型森林地上碳储量的估测精度均显著提高,平均均方根误差由17.23降低到14.3 t·hm －2;直方图匹配后,各区域样地点高值区域低估和低值区域高估现象均有很大程度改善,实测值和估测值间的相关关系明显增强,然而高值地区(碳储量大于20 t·hm －2)出现过高估计现象;经匹配校正后处理的均值、标准差、直方图和累积频率分布图更接近样地实测值,均方根误差也明显降低,高值地区过高估计现象得到很好校正。【结论】森林资源清查数据、遥感数据及 KNN方法相结合逐渐成为区域尺度森林参数空间连续估测的重要手段。同利用光谱值和森林参数建立的回归模型相比,KNN方法能够更多地考虑到森林参数同光谱值之间的非线性依赖关系;但 KNN估测方法除了受距离度量标准、最邻近值 k的大小以及影像波段的选取等因素影响外,还存在如样地坐标和对应的影像光谱值匹配误差、像素级估测结果多呈明显集中分布趋势等问题,使得该方法的应用受到一定限制。本文的研究表明,对这些因素进行合理的校正,将更有利于区域尺度森林参数的精确估计和反演。     

The effect of orifice shape and the injection pressure on enhancement of the atomization process for pressure-swirl atomizers

The following paper presents results of studies on the atomization process of water in pressure-swirl atomizers with different shapes of orifice using a laser diffraction technique. The effect of the shape of the orifice and the injection pressure on the droplet size distributions and mean droplet diameters was analyzed. The results obtained showed that droplet size decreases with an increasing injection pressure and droplet size distributions are expanded. The smallest droplet diameters were obtained by atomization in profiled and conical shaped orifice atomizers, however the largest droplet diameters were achieved for plain orifice atomizers. The analysis of the interfacial surface produced by the atomizers has shown that the greater enhancement of the atomization process had occurred using profiled and conical shaped orifice atomizers compared with plain orifice atomizers. The correlation for the Sauter mean diameter as a function of the Reynolds number and the ratio of orifice length to diameter has been proposed. The paper is related to agricultural water management. The data obtained are important from the point of view of the design of atomizers, agricultural treatments and crop protection.     

改进田间杂草识别图像预处理方法的研究

机器视觉系统采集到的田间杂草图像在分割之前,需要对图像进行预处理来改善图像的视觉效果。为了提高图像的预处理效果,针对杂草图像的特点,在多级加权中值滤波算法的基础上,增加3个不同大小尺寸的二维滤波掩膜,使得在去除噪声的同时尽可能保留了更多的细节特征,从而改善了田间杂草图像的滤波效果。     

基于结构相似度的关中平原旱情空间分布特征

季节性干旱是影响关中平原农业生产最主要的灾害,研究气候变化背景下的关中平原旱情空间分布特征和变化规律具有重要的科学意义与应用价值。基于关中平原2003—2014年每年3月上旬—5月下旬的旬时间尺度条件植被温度指数(VTCI)的干旱监测结果,采用结构相似度(SSIM)研究关中平原VTCI的空间变化特征及其影响因子。结果表明:关中平原VTCI的空间变化特征具有明显的区域差异性和空间异质性,结构性因素是VTCI空间变化的主要影响因子,总体上受水热、地形分布格局等因素的结构性影响,但微地貌、下垫面覆盖的不同及变化和人为因素导致其主体产生变化。当微地貌、下垫面覆盖类型和人为因素成为主导因子时,结构相似度与其高度相关,对其变化的响应非常敏感。与利用某些指数直接作为干旱的指示因子或构建统计模型分析干旱特征相比,引入研究图像的结构信息的参量SSIM实现了定量、准确、形象地描述干旱的空间分布特征和变化规律。     

跌水坎流体动力性噪声模拟试验研

以流体声学相似原理为依据，对三种跌水坎形式进行噪声模拟试验。利用古典声学理论对背景噪声进行分析与修正，评估降噪的效果，确定跌水坎的修改形式。并对修改方案进行不同比尺的试验，对声强进行曲线拟合，绘制趋势线，推求原型噪声的大小，为工程建设提供参考依据。     

一种优化的基于协同过滤的农业信息推荐系统研究

随着信息爆炸而引发的信息过载和资源迷向的问题日趋严重,特别是对于农民来说,对计算机技术的掌握不多。为了向农民及时、方便、高效地提供其感兴趣的农业信息及知识,许多学者都对推荐系统技术进行了研究。为此,首先介绍了目前推荐系统的关键技术,并针对目前推荐系统中存在的用户评价矩阵稀疏问题提出了通过项目聚类来构造出一个无缺失的用户评价矩阵。考虑到农民行为差异性,在进行评价预测之前先对评分进行标准化处理。     

Italian rice varieties: historical data, molecular markers and pedigrees to reveal their genetic relationships

Two molecular marker approaches [amplified fragment length polymorphism (AFLP) and simple sequence repeat (SSR)] were employed to study genomic relationship among 96 rice cultivars. These included most of the best reputed Italian accessions. AFLP produced 461 fragments, 248 (53%) of which were polymorphic, SSR produced four to 11 alleles in the 12 genomic loci investigated. Genomic similarity was estimated independently for the two molecular marker techniques. Both AFLP and SSR dendrograms agree in splitting the cultivars into two main clusters: a small one, comprising four exotic accessions, and a larger one which could be split into four subgroups. These were also analysed on the basis of historical and pedigree information. This is the first report on the application of DNA polymorphism analysis to reveal genomic relationship among cultivated Italian rice germplasm. Results will be useful for breeding programmes.     

Genetic diversity and relationships among pea cultivars revealed by RAPDs and AFLPs

In order to obtain an overview of the genetic diversity present within the set of pea cultivars released in Germany, 21 cultivars were analysed at the DNA level by random amplified polymorphic DNAs (RAPDs) and amplified fragment length polymorphisms (AFLPs), as well as for agronomic traits. Yield of grain cultivars ranged from 2.95 to 3.87 t/ha. Based on the screening of 60 RAPD primers and 32 Eco RI + 3/Mse I+3 AFLP primer combinations, 20 RAPD primers and 11 Eco RI + 3/MseI+ 3 primer combinations generating polymorphic and distinct fragments were chosen for estimation of genetic diversity. Twenty RAPD primers amplified a total of 314 scorable bands ranging from about 262 bp to 1996 bp. Of these, 175 fragments (55.7%) were polymorphic. Based on these data, genetic similarity (GS) was estimated between 0.80 (‘Lisa’ vs.‘Grapis’) and 0.94 (‘Bohatyr’ vs. ‘Sponsor’; mean GS = 0.88). Eleven AFLP primer combinations led to the amplification of 949 scorable fragments ranging from 43 to 805 bp and of these, 462 (48.7%) were polymorphic. Genetic similarity based on AFLPs was calculated between 0.85 (‘Lisa’ vs.‘Laser’) and 0.94 (‘Bohatyr’ vs. ‘Sponsor’, mean GS = 0.90). Correlation of genetic similarity estimated on RAPDs and AFLPs was estimated at r = 0.79** using Spearman's rank correlation coefficient and at r = 0.84 by the Mantel test, respectively. UPGMA cluster analysis carried out on these data separately for RAPDs and AFLPs and on the combined data reflected, to some extent, pedigree relationships and cophenetic correlations (r = 0.89 for RAPDs, r = 0.88 for AFLPs, and r = 0.93 RAPDs + AFLPs) indicate a good fit of respective clusters to genetic similarity data. The correlation of cluster analyses to pedigree information and the impact on parental genotype selection is discussed.     

Reasoning Method Based on Motorcycle Design

According to the needs of setting up a professional system based on motorcycle design,the CBR(case based reasoning) and RBR(rule based reasoning) methods are introduced.After discussing the difference between the two kinds of reasoning,a method of integrating them into the motorcycle design flow is put forward to realize the design reusable technology and improve the design efficiency,by using the CBR in the step of selecting the similar motorcycle and the RBR in project's modification according to correlative rules.     

Weighting Coefficient Optimization Unification of Lexicographic Method in Fuzzy Multi -objective Optimization Design

The Lexicographic method has subjective and empirical character,so a weighting coefficient is introduced by considering the importance and proportion of each individual object,the preferable result of every individual object can be confirmed by comparing the degree of similarity between the perfect one and the ecumenic one and the Fuzzy optimization will be perfect.The effectiveness of these methods is illustrated by a fuzzy optimal design of the three-bar truss.