中文摘要：

【目的】数字图像技术可以用来快速无损地预测冬油菜的氮素营养,建立标准化的拍摄参数获取方法,可为不同作物、不同型号相机间结果的互用提供依据。【方法】以冬油菜为试验材料,设置不同氮肥水平（N：0、90、180、270和360 kg/hm^2）田间试验,于苗期（移栽后79～83天）,利用数码相机（NikonD7000,1620万像素）以不同光照强度、时间、高度、角度、照片像素尺寸和拍摄模式进行拍摄,并以不同储存格式进行保存。比较了不同拍摄条件下获取的冠层数字图像信息差异显著性,同时测定植株地上部生物量、叶片氮浓度和叶绿素浓度,分析冠层图像数字化指标（红光标准化值,NRI）与测定的氮素营养指标之间的相关性。【结果】晴天与阴天获取的冠层NRI均可较好的表达冬油菜冠层数字信息,晴天效果稍好于阴天;中午获取的冠层NRI与叶片氮浓度相关系数为–0.802＊＊,优于上午和下午;1.5、2.0和2.5 m拍摄高度获取的冠层NRI差异不显著;30°、60°和90°拍摄角度下获取的冠层NRI与地上部生物量、叶片氮浓度和叶绿素浓度的相关性均达到极显著水平,30°～60°时获取冠层数码信息可操作性较强;拍摄模式为自动曝光模式获取冠层NRI与叶片氮浓度的相关系数为–0.802＊＊,高于其他拍摄模式;三种照片像素尺寸（4928×3264、3696×2448和2464×1632）下获取的冠层NRI差异不大;储存格式为JPEG精细格式时获取的冠层NRI优于其他储存格式。【结论】综合分析认为,数字图像技术估测冬油菜氮素营养指标最佳操作范围为拍摄时间为晴天太阳高度角相对较大的中午;拍摄高度为近地面均可;拍摄角度为易于操作的30°～60°;相机拍摄模式为自动曝光模式;照片像素尺寸为图片相对较为清晰的高分辨率;存储格式为压缩格式且占用空间较小的JPEG精细格式。本研究为无人机低空遥感的氮素营养无损

英文摘要：

【Objectives】 Digital image technique is useful for rapid and non-destructive prediction of nitrogen concentration of winter rapeseed. The establishment of the standardization of the shooting parameters can provide the basis for the interaction of different crops and camera models. 【Methods】 Winter rapeseed was used infield experiment with five nitrogen application rates（N： 0, 90, 180, 270 and 360 kg/hm2）. The canopy pictures of winter rapeseed were obtained in different light intensities, times, heights, angles, pixels, modes and image formats by using digital camera（Nikon-D7000, 1620 million pixels） at the seedling stage（79–83 days after transplanting）. Significant analyses of canopy digital normalized redness intensity（NRI） were also made in different imagery modes. Other conventional diagnosis parameters, such as aboveground biomass, leaf nitrogen and leaf chlorophyll concentration were determined to reveal their correlations to NRI. 【Results】 Canopy NRI acquired either at sunny or cloudy days could interpret digital information of canopy well, however, it was slightly better in sunny days than cloudy days. The correlation coefficient between NRI and leaf nitrogen concentration was – 0.802＊＊ at noon, better than those in the morning and afternoon. No significant difference was observed among three image height of 1.5, 2.0 and 2.5 m. It showed very prominent relationships between NRI obtained in three image angles（30°, 60°and 90°） and aboveground biomass, leaf nitrogen concentration and chlorophyll concentration, however, it was more convenient to get canopy NRI at 30°–60°. The correlation coefficient between NRI and leaf nitrogen concentration was – 0.802＊＊ at auto exposure mode, higher than other exposal models. No significant difference was observed among three photo pixels（4928 × 3264, 3696 ×2448 and 2464 × 1632）. NRI in JPEG fine format was superior to other image formats. 【Conclusions】Overall, the optimal operating range of digital image te