【作者】 张敏；

【导师】 潘存德；

【作者基本信息】 新疆农业大学 ， 森林培育学， 2021， 硕士

【摘要】 ‘新榛1号’(Corylus’Xinzhen 1’)具有高产、种仁饱满、出仁率高、抗寒、适应性强等特性,是近几年在新疆在北疆地区和南疆冷凉山区推广栽植的平欧杂种榛(C.heterophylla×C.avellana)优良品种。研究建立叶片矿质营养元素含量光谱反演模型,可为其高效、适时的以叶片营养元素含量为指示的树体营养监测提供技术途径。鉴于此,本文以盛果期‘新榛1号’为研究对象,基于肥料效应田间实验,以便携式光谱分析仪(Uni Spec-SC)作为叶片光谱反射率的测定仪器,采用田间活体叶片光谱反射率测定、叶样采集与室内矿质营养元素含量化学测定相结合的方法,在筛选果实4个生育时期[果实坐果期(FSP)、果实速生生长期(FRGP)、果实脂化期(FFCP)和果实近成熟期(FNMP)]表征叶片矿质营养元素[氮(N)、磷(P)、钾(K)、钙(Ca)、镁(Mg)、铁(Fe)、锰(Mn)、铜(Cu)、锌(Zn)]含量的有效光谱特征参量和有效光谱敏感波段的基础上,通过分析果实不同生育时期建立叶片矿质营养元素含量光谱反演统计模型可采用的数学函数关系,采用参数的最小二乘估计方法建立‘新榛1号’果实4个生育时期叶片矿质营养元素含量光谱反演统计模型。所得主要研究结果如下:(1)果实4个生育时期叶片N、P、K营养元素存在表征其各自含量的有效光谱特征参量。表征N元素最有效的光谱特征参量在果实4个生育时期分别为绿色归一化差值指数[(NIR-Green)/(NIR+Green)]、红色归一化差值指数[(NIR-Red)/(NIR+Red)]、红边面积(SD_r)和SD_r;表征P元素最有效的光谱特征参量在果实4个生育时期分别为红边蓝边面积比值(SD_r/SD_b)、蓝边面积(SD_b)、红谷反射率(R_o)和绿峰反射率(R_g);表征K元素最有效的光谱特征参量在果实4个生育时期分别为红边蓝边面积归一化值[(SD_r-SD_b)/(SD_r+SD_b)]、绿色比值指数(RNIR/Green)、R_o和R_o。(2)果实4个生育时期叶片Ca、Mg、Fe、Mn、Cu、Zn营养元素存在表征其各自含量的有效光谱敏感波段。叶片Ca元素在FSP为327、334、339、354、372、395、447、456、496、563、809、814、823、831、986、1 046和1 059 nm;在FRGP为351、359、378、409、471、774、775、845、866、894、923、979、990、1 014、1 031和1091 nm;在FFCP为388、440和993 nm;在FNMP为365、485、508、536、538、571、603、607、795、835、885、978、991、1 040、1 072、1 078和1 097 nm。叶片Mg元素在FSP为321、384、425、469、669、785、795、853、894、922、970、980、1 013、1 038和1 073 nm;在FRGP为317、335、360、379、485、846、902、948、974、981、984、1 015、1 036、1 062、1 075和1 092 nm;在FFCP为467、517和1 095 nm;在FNMP为332、346、366、449、470、546、589、643、701、882、902、903、918、994、1 041、1 045和1 064 nm。叶片Fe元素在FSP为315、341、398、429、557、581、815、830、832、941、955、974、978、1 045、1 073、1 086和1 112 nm;在FRGP为347、352、395、456、469、474、574、773、784、850、853、928、959、969、1 110和1 124 nm;在FFCP为820、845、1 085和1 102 nm;在FNMP为391、439、456、509、522、545、653、678、767、826、834、856、1 035、1 050、1 070、1 098和1 109 nm。叶片Mn元素在FSP为341、358、440、484、525、534、633、754、802、827、840、891、965、975、1 016和1 045 nm;在FRGP为341、358、440、484、525、534、633、754、802、827、840、891、965、975、1 016和1 045 nm;在FFCP为350、778、851、868和1 048 nm;在FNMP为369、387、393、494、522、581、599、609、629、653、859、922、974、1 016和1 085 nm。叶片Cu元素在FSP为351、353、366、402、420、675、779、836、870、948、1 027、1 072、1 120和1 129 nm;在FRGP为331、340、396、441、468、769、837、878、919、1 004、1 016、1 033、1 045和1077 nm;在FFCP为378、933和992 nm;在FNMP为314、505、506、569、623、640、674、775、836、1 033、1 037、1 053、1 060和1 116 nm。叶片Zn元素在FSP为342、364、422、552、560、585、596、601、623、641、862、880、1 076、1 098、1 099和1 112 nm;在FRGP为339、362、365、382、576、639、821、876、942、1 010、1 032、1 049、1 056、1 092、1 099和1 112 nm;在FFCP为450和1 061 nm;在FNMP为316、378、396、452、491、555、591、609、774、805、876、908、923、1 040、1 089和1095 nm。(3)建立叶片N、P、K矿质营养元素含量光谱反演统计模型,可采用三次函数,并以有效光谱特征参量作为自变量。(4)建立叶片Ca、Mg、Fe、Mn、Cu、Zn矿质营养元素含量光谱反演统计模型,可采用线性函数关系,并以有效光谱敏感波段光谱反射率的一阶微分为自变量。更多还原

【Abstract】 Corylus’Xinzhen 1’is an excellent(C.heterophylla×C.avellana)which has been popularized and planted in northern and southern Xinjiang in recent years due to its high yield,full kernel,high kernel yield,cold resistance and strong adaptability.Some studies have established a spectral inversion model of leaf mineral nutrient content,which can provide a technical way for efficient and timely tree nutrient monitoring based on leaf nutrient content.Therefore,based on the field experiment of fertilizer effect and taking Corylus’Xinzhen 1’in the full fruit stage as the research object,this paper used a portable spectrometric analyzer(Uni Spec-SC)as the instrument to measure the spectral reflectance of leaves.Using the method of combining spectral reflectance measurement of living leaves in field,leaf sample collection and chemical determination of mineral nutrient content in laboratory,we screened the effective spectral characteristic parameters and the effective spectral sensitive bands for the content of mineral nutrients in fruit leaves at four growth periods In addition,based on the analysis of different growth periods of fruits(fruit setting period(FSP),fruit rappid growth period(FRGP),fruit fat change period(FFCP)and fruit near mature period(FNMP),this paper establishes the mathematics function relations that can be used in the spectral inversion statistical model of leaf mineral nutrient content(N,P,K,Ca,Mg,Fe,Mn,Cu,Zn).Furthermore,the least square estimation method of parameters was used to establish the spectral inversion statistical model of leaf mineral nutrient content of Corylus’Xinzhen 1’fruits in four growth periods.The main findings were as follows:(1)There were effective spectral characteristic parameters for the contents of N,P and K nutrient elements in leaves of fruit at four growth periods.The most effective spectral characteristic parameters of N were(NIR-Green)/(NIR+Green),(NIR-Red)/(NIR+Red),SD_rand SD_r;The most effective spectral characteristic parameters of P were(SD_r/SD_b),SD_b,R_oand R_g;The most effective spectral characteristic parameters of K were(SDr-SDb)/(SDr+SDb),(RNIR/Green),R_o and R_o.(2)There were effective spectral sensitive bands of Ca,Mg,Fe,Mn,Cu and Zn in leaves of fruit at four growth periods.Leaf Ca elements at FSP were 327,334,339,354,372,395,447,456,496,563,809,814,823,831,986,1 046 and 1 059 nm;351,359,378,409,471,774,775,845,866,894,923,979,990,1 014,1 031 and 1 091 nm in FRGP;388,440 and993 nm in FFCP;365,485,508,536,538,571,603,607,795,835,885,978,991,1 040,1072,1 078 and 1 097 nm in FNMP.Leaf Mg elements at FSP were 321,384,425,469,669,785,795,853,894,922,970,980,1 013,1 038 and 1 073 nm;317,335,360,379,485,846,902,948,974,981,984,1 015,1 036,1 062,1 075 and 1 092 nm in FRGP;467,517 and 1095 nm in FFCP;332,346,366,449,470,546,589,643,701,882,902,903,918,994,1 041,1 045 and 1064 nm in FNMP.Leaf Fe elements at FSP were 315,341,398,429,557,581,815,830,832,941,955,974,978,1 045,1 073,1 086 and 1 112 nm;347,352,395,456,469,474,574,773,784,850,853,928,959,969,1 110 and 1 124 nm in FRGP;820,845,1 085and 1 102 nm in FFCP;391,439,456,509,522,545,653,678,767,826,834,856,1 035,1050,1 070,1 098 and 1 109 nm in FNMP.Leaf Mn elements at FSP were 341,358,440,484,525,534,633,754,802,827,840,891,965,975,1 016 and 1 045 nm;341,358,440,484,525,534,633,754,802,827,840,891,965,975,1 016 and 1 045 nm in FRGP;350,778,851,868 and 1 048 nm in FFCP;369,387,393,494,522,581,599,609,629,653,859,922,974,1 016 and 1 085 in FNMP.Leaf Cu elements at FSP were 351,353,366,402,420,675,779,836,870,948,1 027,1 072,1 120 and 1 129 nm;331,340,396,441,468,769,837,878,919,1 004,1 016,1 033,1 045 and 1 077 nm in FRGP;378,933 and 992 nm in FFCP;314,505,506,569,623,640,674,775,836,1 033,1 037,1 053,1 060 and 1 116 nm in FNMP.Leaf Zn elements at FSP were 342,364,422,552,560,585,596,601,623,641,862,880,1076,1 098,1 099 and 1 112 nm;339,362,365,382,576,639,821,876,942,1 010,1 032,1049,1 056,1 092,1 099 and 1 112 nm in FRGP;450 and 1 061nm in FFCP;316,378,396,452,491,555,591,609,774,805,876,908,923,1 040,1 089 and 1 095 in FNMP.(3)The spectral inversion statistical model of N,P,K mineral nutrient contents in leaves was established,which could use cubic function and effective spectral characteristic parameters as independent variables.(4)We established a spectral inversion statistical model for the contents of Ca,Mg,Fe,Mn,Cu and Zn mineral nutrients in leaves,which can adopt a linear function relationship and take the first-order differential of the spectral reflectance of the effective spectral sensitive bands as the independent variable.更多还原