一种高效低功耗JPEG2000小波变换器的设计

提出了一种高效低功耗的JPEG2000小波变换器的硬件结构,减小了二维离散小波变换器(2-D DWT)中暂存器的大小,降低了存储器的平均访问带宽,有效减小了芯片的面积和功耗.整个设计通过了FPGA验证,并采用HJTC 0.18μm工艺库进行了芯片设计.试验结果表明,片内暂存器的面积和功耗比普通方法分别减小了54.5%和61.4%.     

棉花高光谱及其红边特征(I)

通过大田和室内试验,测定了2个品种的棉花冠层、完全展开倒1、3叶在不同时期的高光谱反射率及对应叶片的叶绿素、类胡萝卜素含量。结果表明:随发育期推移,棉花冠层光谱反射率在可见光范围降低,在近红外区域增高;叶片背面光谱反射率略高于正面,透射率小于反射率;叶面积指数、鲜叶重和干叶重与冠层反射光谱变量ρ800 ρ550、ρ800 ρ680、ρ680 ρ570之间存在显著相关;叶片叶绿素和类胡萝卜素浓度与其反射光谱变量ρ680 ρ570、ρ673 ρ640、ρ680 ρ550、PSSRa、PSNDa、RCh之间也呈显著相关。     

Using colour image analysis for quantitative assessment of powdery mildew on cucumber

Summary Visual assessment, black & white image analysis and colour image analysis of the severity of powdery mildew (Sphaerotheca fuliginea) infection on leaves of cucumber (Cucumis sativus) were compared. The extent of infection with powdery mildew was expressed as the number of spores per cm². Statistical treatment of the data showed colour image analysis to be superior to the other two methods of evaluation. Major veins, light reflections and variations in the colour of leaves can degrade the accuracy of colour image analysis somewhat. These problems are, however, surmountable. Colour image analysis qualifies as a valuable new tool for accurate, objective, reproducible and detailed quantitative assessment of disease on leaves of plants.Abbreviations H hue - I intensity - IA image analysis - PM powdery mildew - S saturation     

Real-Time Self-Photographing Technology in the Virtual Environment

As a new technology towards 21st century, the occurs of the virtual reality technology is a result of computer graphic, human machine interface technology, sensor technology and artificial intelligence. The real time self photographing technology based on the virtual environment is a typical application of this virtual reality technique. A series of things has been discussed and studied ,such as visual reality,video intercept and so on,and the real time self photographing system based on the technology of blue background image extracting. A new arithmetic is emphasized which is produced during study and exploitation.It is image fusion technique based on the contrast appreciate characteristic.     

Color Image Compression Method Based on Incomplete Scanning Method

In view of color image characteristic, the paper improves the zero algorithm in the application of concrete algorithm, draws a new algorithm-incomplete scanning method, which is applied to color image compression. The authorsconvert color image from the RGB space into luminance-chrominance YUV space, then compress the converted image to encode/decode. Theoretical analysis and simulation experiment indicate that the compression ratio and encoding/decoding speed are increased, at the sase time the quality of restored image is guaranteed.     

Index System of Sub-Microstructure Analysis System of Asphalt Concrete(MASAC)

Based on the study of image processing algorithm,the Sub-Microstructure Analysis System of Asphalt Concrete(MASAC) suitable for Windows Operating System was specially programmed.In this paper the index system of MASAC is chiefly illustrated and the functions of a part of indexes are briefly introduced.The index system is classified into 5 types,i.e.characteristics of aggregate particles,arrangement of aggregate particles,contact among aggregate particles,air voids between aggregate particles and distribution of aggregate particles.By analysis for the microstructure of asphalt concrete,it can be used to investigate the internal relationship between asphalt concrete microstructure and macro-performance,and evaluate the asphalt mixture homogeneity.     

棉花高光谱及其红边特征(Ⅱ)

通过大田和室内试验,测定了2个棉花品种的冠层、完全展开倒1、3叶在不同时期的高光谱反射率及对应叶片的叶绿素、类胡萝卜素含量。结果表明:棉花冠层光谱红边具有“双峰”和“红边平台”现象,且红边位置λ_(red)位于695～720nm之间,红边幅值Dλ_(red)和红边面积S_(red)有“红移”和“蓝移”现象;叶面积指数、鲜叶重和干叶重与冠层光谱红边参数λ_(red)、Dλ(red)、S_(red)之间存在显著相关,叶片叶绿素和类胡萝卜素含量与其反射光谱的λ_(red)、Dλ(red)、S_(red)也有显著相关。     

基于地质统计学影像纹理的石河子地区化控期棉花长势监测

化学调控是棉花生长期间的一个重要环节，在较大区域尺度上，化控常常整齐划一，依固定的量而行，忽视其普遍存在的空间差异，结果导致某些棉田投入过量或投入不足，在精细农业所倡导的基于特定场的农业投入这一框架下，为辅助决策者制定有针对性的棉田化控方案，该文提出了应用地质统计学影像纹理进行对化控棉花长势监测，通过借助地质统计学方法在分析空间结构变化上的优势，强化遥感影像上不同长势棉田间的差别，提高样本选择的分离度和长势空间分类精度。实验结果表明，运用变异函数纹理结合光谱波段的最大似然分类方法能够很好地界定棉花长势的不同类型，总分类精度达90.53%，证明了应用地质统计学影像纹理方法在针对化控的棉花长势监测上的有效性。     

模板法自动提取遥感图像耕地变化信息的研究

该文在多时相遥感图像中采用主成份差异法提取耕地变化信息，研究的关键在于差值前如何进行图像匹配、如何自动提取变化信息以及差值后如何确定变化阈值。在逐像元判断是否变化时，根据计算机视觉的数据处理思想提出一种采用(2n+1)×(2n+1)判断模板进行判断的方法，这种方法可以大幅度地改善单像元判断的效果。经试验，提取耕地变化信息的准确程度达85％以上，可以替代人工判读的工作。     

TM影像中大棚菜地信息提取技术研究

大棚菜地在我国北方呈不断发展之势,快速、准确地获取其面积及分布状况,是蔬菜业健康发展和科学管理的客观需求。为探索快速准确的大棚菜地信息提取技术方法,以山东省青州市为例,采用TM数字图像,深入分析各主要地物光谱特征的基础上,构建了能够突出反映大棚菜地信息的大棚菜地指数模型,同时通过对图像的分区、掩膜处理,以及研究区背景数据及专家知识的应用,准确提取了大棚菜地信息。结果显示:青州市2002年4月14日的大棚菜地面积为21375.4hm2,集中分布于青州东北部平原区的谭坊、郑母、黄楼等乡镇。本研究技术方法对于大棚菜地信息的遥感提取有积极参考价值,为研究区及山东蔬菜业的合理发展提供重要技术支撑。