朴素贝叶斯算法在蝴蝶兰病害分类中的应用研究

蝴蝶兰属于热带气生兰,对生长环境的温湿度非常敏感,温湿度过高或过低不仅对蝴蝶兰生长不利而且易引发多种病害。为此,利用传感器采集的蝴蝶兰生长环境空气温湿度数据,运用朴素贝叶斯算法对其进行分类,以此判断蝴蝶兰的健康状况,达到病害分类的目的。实验表明,该方法的平均分类准确率达到81%,与其它统计学预报方法相比,具有简单、易行的优点。     

基于RAGA的PPC模型在黑土区玉米调亏灌溉评价中的应用

调亏灌溉作为一种行之有效的节水灌溉方式,对黑土区农业水资源的可持续高效利用具有重要意义。而调亏灌溉方案的优劣取决于它的综合效果,难以直观评价。以2013年在黑龙江省红星农场进行的玉米调亏灌溉技术效应的试验研究为基础,选取了8个评价指标,采用投影寻踪分类模型(PPC),利用基于实数编码的加速遗传算法(RAGA)优化其投影方向,根据投影函数值得到各评价指标对综合评价的贡献大小对不同调亏处理的各项指标进行排序。结果表明:苗期水分调亏程度为田间持水量60%的处理为最佳调亏灌溉方案,产量和水分利用效率比适宜调亏灌溉处理分别提高了2.11%和6.42%。研究结果可为黑土区玉米灌溉管理提供科学依据。     

The genetic diversities of 4 species of subg. Lithocerasus (Prunus, Rosaceae) revealed by RAPD analysis

RAPD analysis was applied to reveal the genetic diversities of 4 speciesof subg. Lithocerasus within the genus Prunus using 40accessions representing the subgenera Prunophora, Amygdalus,Lithocerasus and Cerasus. The accession of subg. Lithocerasus are phenotypically similar to members of subgenus Cerasus but different with them in interspecific crossing tests and isozymeprofiles. Two major clusters, `Prunophora and Amygdalus group' and `Cerasus group' were constructed in the phenogram. We revealthat the examined 4 species of subg. Lithocerasus species; P.tomentosa Thunb., P. japonica Thunb., P. glandulosa Thunb.and P. besseyi Bailey, were genetically closer to the members ofsubgenera Prunophora and Amygdalus than to subg. Cerasus.     

Classification of Azolla spp., section Azolla

Summary Azolla accessions (section Azolla) from the germplasm collections of the International Rice Research Institute and Washington State University were fingerprinted and classified by enzyme electrophoresis and leaf trichome morphology. A. filiculoides was enzymatically distinctive and also reliably identified by its prominent one-celled trichomes. Neotropical accessions labelled as A. filiculoides proved to be members of other species. Two groups of isolates were designated A. rubra, but those from Japan were identified as A. filiculoides. The A. rubra of Australia-New Zealand was biochemically unique and possessed less protuberant trichomes than A. filiculoides. A. microphylla, A. mexicana, and A. caroliniana were phenetically similar, but a. microphylla was identifiable from the others in the banding patterns of certain enzymes. A. mexicana and A. caroliniana were closely related enzymatically. The two-celled leaf trichomes of these three species were similar in size and shape.     

MEASURING PROBABILITY OF SKY BRIGHTNESS IN CLASSIFICATION BY THE AVERAGE INSOLATION

According to the daily change law from curves for percentage insolation,this paper presents a new formula of the average insolation in the zones where the insolation is more than 70%. By means of the hourly observing data during two years for solar direct ratiation, cloudage and sunshine condition in seven zones different lighting climate,considering the vaper pressure and the altitude present the function of percentage insolation and the probability of four sky brightness classification,this formula has a certain precision to be proved by observation.     

BP神经网络在储粮害虫分类中的应用研究

综合利用计算机视觉技术和BP神经网络技术,实现对粮仓害虫的无损检测.通过对粮仓害虫图像的CCD图像预处理,获取了几何特征和不变矩等15个特征参数,并通过优化选取其中七个参数输入神经网络进行训练.仿真结果表明训练网络对粮仓四类常见害虫的识别率达到了85%,得到了较好的识别结果.     

RFLP characterization of Indian Musa germplasm for clonal identification and classification

Summary Nineteen single-copy clones isolated from a PstI genomic library (cv. Maiden Plantain), and eight Vigna chloroplast DNA clones were used to probe total genomic DNA digests of 57 genotypes of Musa from India. The 19 genomic clones detected a total of 107 polymorphisms among the 57 genotypes. Principal coordinates and phenetic analyses of these data placed cultivars and species into distinct groups that were in general agreement with a previously published RAPD-based classification of these same plant materials. The 107 polymorphisms were sufficient to differentiate each clone from every other clone. Heterologous Vigna chloroplast DNA probes were used to characterize the cytoplasm of Musa cultivars and species. PCO analysis of these RFLPs were detected both within and between the generally recognized genome groups, indicating multiple hybridization pathways in the origin of hybrid clones. Data presented demonstrate that RFLPs are sufficiently abundant to classify Musa germplasm and that genetic relationships among Musa cultivars, based upon RFLP data, are in general agreement with relationships determined by analysis of morphology and RAPDs.     

Origin,classification, variation and distribution of lima bean (Phaseolus lunatus L.) in the light of Cuban material

Summary Recent collecting activities have shown that Cuba is rich in plant genetic resources of lima beans. Seed characters of 173 Cuban accessions were measured. Principal component analysis and discriminant factor analysis were used to compare the accessions, which could be classified into seven groups: wild, weed, Potato, Potato-Sieva, Sieva, Sieva-Big Lima, Big Lima. The Sieva type is the principal group of Cuba. It has arrived to Cuba in pre-Columbian times via the Antilles arc. After the Conquest the Potato type was introduced from Central America. Introgressive hybridization with the Sieva group led to the intermediate Potato-Sieva type, whereas the Sieva-Big Lima type resulted from selection toward larger seeds within the Sieva group. The Big Lima type was introduced relatively late. Cuba occupies a key position in the distribution of Lima beans within the Americas and also to the Old World.     

中国西部典型区生态环境本底遥感调查

1997年我国实施了西部大开发战略,西部地区的生态环境问题再次引起多方面的关注。“西部金睛行动计划”的实施旨在通过具有代表意义的7个典型区同步监测西部10年(1990～2000年)来的生态环境变化情况,为国家西部大开发战略提供科学的决策支持。项目采用遥感与地理信息系统技术相结合的方法,按照8个统一的原则,制订了生态环境遥感监测技术指南,对这些地区在1990年和2000年的生态环境变化进行了动态监测。监测结果表明,这10年来,由于人类活动加强,农业用地、城市用地大量增加,挤占了自然/半自然植被区,草场过牧引起草场退化,致使典型区生态环境仍处于逐渐恶化的状态。