肖春安,蔡甲冰,常宏芳,張敬曉,許 迪
考慮作物生長(zhǎng)狀態(tài)的農(nóng)田表面溫度數(shù)據(jù)精量甄別與區(qū)分
肖春安,蔡甲冰※,常宏芳,張敬曉,許 迪
(1. 中國(guó)水利水電科學(xué)研究院,流域水循環(huán)模擬與調(diào)控國(guó)家重點(diǎn)實(shí)驗(yàn)室,北京,100038;2. 國(guó)家節(jié)水灌溉北京工程技術(shù)研究中心,北京,100048)
農(nóng)田表面溫度是土壤、作物和大氣之間進(jìn)行水/熱交換傳輸?shù)闹匾獏?shù),也是灌區(qū)遙感反演模型的重要參量。在利用熱紅外傳感器連續(xù)獲取農(nóng)田表面溫度數(shù)據(jù)時(shí),由于作物的生長(zhǎng)發(fā)育處于動(dòng)態(tài)變化中,農(nóng)田表面溫度數(shù)據(jù)往往混合了作物冠層溫度和土壤表面溫度。為精準(zhǔn)甄別和區(qū)分田間海量監(jiān)測(cè)數(shù)據(jù),該研究結(jié)合Logistic作物生長(zhǎng)模型,通過考慮作物生長(zhǎng)狀態(tài)指標(biāo)葉面積指數(shù)(Leaf Area Index,LAI)和作物冠層高度及其關(guān)鍵節(jié)點(diǎn),構(gòu)建了農(nóng)田表面溫度監(jiān)測(cè)數(shù)據(jù)的甄別算法。以內(nèi)蒙古永濟(jì)試驗(yàn)站玉米和向日葵實(shí)測(cè)數(shù)據(jù)對(duì)算法進(jìn)行驗(yàn)證,并利用解放閘灌域和吉林省長(zhǎng)春試驗(yàn)站的玉米和向日葵田間觀測(cè)數(shù)據(jù)進(jìn)行校核。結(jié)果表明:考慮LAI和作物冠層高度并利用Logistic模型模擬的關(guān)鍵節(jié)點(diǎn)來(lái)建立甄別算法,能夠?yàn)檗r(nóng)田稀疏植被表面溫度數(shù)據(jù)甄別提供高效判定。與人工測(cè)量值對(duì)比,冠層溫度優(yōu)化幅度在10 個(gè)百分點(diǎn)左右(相對(duì)誤差),土壤表面溫度優(yōu)化幅度超過5個(gè)百分點(diǎn);甄別方法可以明顯提升冠層和土壤表面溫度的獲取精度。甄別算法中校正因子數(shù)值需根據(jù)作物種植密度及LAI確定,其中玉米校正因子選擇作物冠層溫度校正因子0.9,土壤表面溫度校正因子1.1;向日葵校正因子以葉面積指數(shù)最大值4為基礎(chǔ),選取冠層溫度校正因子0.7,土壤表面溫度校正因子1.2;在不同地區(qū)應(yīng)用時(shí),向日葵葉面積指數(shù)最大值每增加1,推薦冠層溫度校正因子調(diào)高0.35,土壤表面溫度校正因子調(diào)低0.18。研究結(jié)果可為精量灌溉提供技術(shù)支撐,提高了農(nóng)田監(jiān)測(cè)數(shù)據(jù)的性能,為無(wú)人機(jī)遙感和衛(wèi)星遙感數(shù)據(jù)的精量甄別提供算法和驗(yàn)證。
溫度;傳感器;土壤;冠層;數(shù)據(jù)甄別;Logistic模型
隨著物聯(lián)網(wǎng)、云平臺(tái)等無(wú)線通信技術(shù)的進(jìn)步與發(fā)展,農(nóng)業(yè)生產(chǎn)管理也逐步走向精準(zhǔn)高效;精準(zhǔn)農(nóng)業(yè)能夠結(jié)合作物需求,根據(jù)農(nóng)田環(huán)境實(shí)時(shí)信息精準(zhǔn)并有針對(duì)性的對(duì)作物生長(zhǎng)進(jìn)行管理,為實(shí)現(xiàn)作物高效生產(chǎn)提供重要保障[1-3]。其中的精量灌溉依托于農(nóng)情監(jiān)測(cè)數(shù)據(jù),對(duì)作物自身水分信息、農(nóng)田小氣候以及土壤墑情等因素進(jìn)行綜合評(píng)判,進(jìn)而獲得最精準(zhǔn)的信息進(jìn)行灌溉決策、預(yù)報(bào)和管理[4-6]。
借助ZigBee、Wi-Fi、LoRa、NB-IoT和RFID技術(shù)等多種無(wú)線通信方式,無(wú)線傳感器網(wǎng)絡(luò)能高效準(zhǔn)確地采集和傳輸農(nóng)田環(huán)境信息,遠(yuǎn)程監(jiān)控模塊的增添實(shí)現(xiàn)了農(nóng)田小氣候與視頻圖像信息參數(shù)采集與傳輸?shù)母叨燃蒣7-10],三維GIS的應(yīng)用可展示大規(guī)模農(nóng)田作業(yè)區(qū)三維虛擬場(chǎng)景[11-12]。Sakamoto等[13]利用數(shù)碼相機(jī)捕捉作物近紅外圖像進(jìn)而連續(xù)監(jiān)測(cè)作物狀態(tài),提供了可用于驗(yàn)證來(lái)自MODIS等衛(wèi)星系統(tǒng)長(zhǎng)時(shí)間序列的地面真實(shí)數(shù)據(jù)的替代方法;Kim等[14]利用集成的低成本、近地表遙感系統(tǒng)對(duì)植被指數(shù)、光合有效輻射和葉面積指數(shù)進(jìn)行植被冠層動(dòng)態(tài)的持續(xù)監(jiān)測(cè);Xiang等[15]采用高時(shí)空分辨率的遙感系統(tǒng)自動(dòng)捕獲田間多光譜圖像進(jìn)而可對(duì)作物冠層反射率校準(zhǔn)。農(nóng)田遠(yuǎn)程墑情、作物冠層溫度監(jiān)測(cè)和農(nóng)田環(huán)境信息的相結(jié)合,可為農(nóng)田綜合灌溉決策提供及時(shí)、準(zhǔn)確的數(shù)據(jù)[16-18]。
農(nóng)田表面溫度常應(yīng)用在估算農(nóng)田蒸散發(fā)方法和模型中[19-21]。作物冠層溫度是研究土壤、作物和大氣之間進(jìn)行水/熱交換傳輸?shù)闹匾獏?shù),可利用其與氣溫的差值來(lái)判斷作物缺水狀況[22-24];準(zhǔn)確的作物冠層溫度數(shù)據(jù)也可作為抗旱和耐熱作物品種選育的重要依據(jù)[25]。熱紅外傳感器測(cè)溫具有較高的穩(wěn)定性,因其尺度擴(kuò)展和適應(yīng)性,在農(nóng)田尺度和區(qū)域尺度都能廣泛使用[26-28]??衫脽峒t外傳感器自動(dòng)監(jiān)測(cè)系統(tǒng)對(duì)農(nóng)田下墊面進(jìn)行連續(xù)多點(diǎn)掃描[29];因作物生長(zhǎng)發(fā)育的進(jìn)程,以及行距株距的存在,尤其是在作物生育早期所獲得的溫度數(shù)據(jù)可能會(huì)包含田間作物冠層溫度和土壤表面溫度。目前常使用監(jiān)測(cè)溫度數(shù)據(jù)的均值,簡(jiǎn)化了數(shù)據(jù)處理過程,但混合數(shù)據(jù)可能會(huì)影響精細(xì)農(nóng)田灌溉模型計(jì)算精度。如常用的農(nóng)田雙源蒸散發(fā)模型分別估算作物蒸騰和土壤蒸發(fā),其關(guān)鍵點(diǎn)就是要準(zhǔn)確獲取作物冠層和土壤表面的溫度及相關(guān)參量[30-33]。因此對(duì)農(nóng)田表面溫度進(jìn)行甄別與區(qū)分以獲得翔實(shí)的溫度數(shù)據(jù),將有助于提升模型估算精度并得到有效應(yīng)用。另外,通過對(duì)農(nóng)田表面溫度進(jìn)行有效甄別和區(qū)分,亦為研究農(nóng)田水熱梯度變化規(guī)律提供了同步的作物冠層溫度和土壤表面溫度數(shù)據(jù),可保證農(nóng)田溫度時(shí)空差異性研究的精準(zhǔn)可靠。
為有效甄別和區(qū)分農(nóng)田表面溫度數(shù)據(jù)并提高數(shù)據(jù)的有效性及可利用性,本文利用作物生長(zhǎng)指標(biāo)構(gòu)建一種溫度數(shù)據(jù)甄別方法,對(duì)玉米和向日葵2種作物的農(nóng)田表面溫度監(jiān)測(cè)數(shù)據(jù)進(jìn)行優(yōu)化處理。以內(nèi)蒙古河套灌區(qū)永濟(jì)灌域的田間試驗(yàn)觀測(cè)數(shù)據(jù)為例,綜合評(píng)判甄別區(qū)分后所得作物冠層溫度和土壤表面溫度數(shù)據(jù)的有效性;進(jìn)一步應(yīng)用河套灌區(qū)解放閘灌域和吉林省長(zhǎng)春試驗(yàn)站的田間實(shí)測(cè)數(shù)據(jù)進(jìn)行評(píng)估驗(yàn)證,探究該甄別方法的適用性,以期為精準(zhǔn)農(nóng)業(yè)和精量灌溉的農(nóng)田實(shí)時(shí)數(shù)據(jù)提供高效管理方法和處理依據(jù),為科學(xué)灌溉的實(shí)施提供重要技術(shù)支撐。
1.1.1 監(jiān)測(cè)系統(tǒng)的布設(shè)
2021年5—9月作物主要生育期內(nèi),在內(nèi)蒙古河套灌區(qū)永濟(jì)試驗(yàn)站(107°16′35″E,40°44′11″N)玉米(科合699)和向日葵(谷豐6號(hào))的田塊中間分別安裝CTMS-On line系統(tǒng),如圖1所示。本年度玉米和向日葵播種日期分別是5月20日和6月12日,出苗日期為5月28日和6月19日,收獲日期分別是10月2日和9月30日。玉米和向日葵的種植密度分別為7.5和4株/m2。CTMS-On line系統(tǒng)通過旋轉(zhuǎn)平臺(tái)驅(qū)動(dòng)懸臂梁末端的熱紅外溫度傳感器,對(duì)田間下墊面進(jìn)行多點(diǎn)掃描,其詳細(xì)工作原理可見參考文獻(xiàn)[18]。如圖1c所示,置于監(jiān)測(cè)系統(tǒng)旋轉(zhuǎn)臂上的熱紅外傳感器每次環(huán)繞中心立柱,按照固定角度均勻掃描下墊面1周,可獲取10個(gè)點(diǎn)位溫度數(shù)據(jù)。掃描系統(tǒng)采集時(shí)間間隔為30 min,同步連續(xù)監(jiān)測(cè)的其他指標(biāo)還包括太陽(yáng)輻射、光合有效輻射、空氣溫/濕度、風(fēng)速、大氣壓強(qiáng),以及作物根區(qū)土壤溫/濕度(作物根區(qū)10、20和40 cm)等。由于作物冠層發(fā)育過程和作物行距株距的存在,掃描點(diǎn)位可能在作物冠層和行間距間的土壤表面(圖1c)。
圖1 玉米和向日葵試驗(yàn)地CTMS-On line監(jiān)測(cè)系統(tǒng)的布設(shè)與手持熱紅外測(cè)溫儀觀測(cè)情況
2016年內(nèi)蒙古河套灌區(qū)解放閘灌域在向日葵和玉米地分別布置CTMS-On line系統(tǒng)(106°43′~107°27′E,40°34′~41°14′N),向日葵和玉米的種植密度分別為5 和7.5株/m2。2018—2019年在吉林省長(zhǎng)春試驗(yàn)站(125°19′8″E,43°38′40″N)玉米地布設(shè)了相同監(jiān)測(cè)系統(tǒng),兩年玉米的種植密度均為8株/m2。以上系統(tǒng)連續(xù)監(jiān)測(cè)所獲得的數(shù)據(jù)用于設(shè)計(jì)算法和模型驗(yàn)證。
1.1.2 其他監(jiān)測(cè)項(xiàng)目
作物出苗后,每2 d于10:00—18:00間隔2 h使用手持熱紅外測(cè)溫儀(ST80+,美國(guó)雷泰公司),在布設(shè)監(jiān)測(cè)系統(tǒng)的玉米和向日葵地定點(diǎn)測(cè)取8~10組作物冠層溫度和土壤表面溫度(圖1d),并對(duì)其進(jìn)行均值化處理以消除測(cè)量主觀誤差和隨機(jī)誤差。此數(shù)據(jù)將用于與CTMS-On line系統(tǒng)監(jiān)測(cè)數(shù)據(jù)的驗(yàn)證和對(duì)比。
在生長(zhǎng)季內(nèi),分別在玉米和向日葵地固定選取3株,每隔7~10 d測(cè)量作物冠層高度()、葉長(zhǎng)和葉寬等作物生態(tài)指標(biāo)。葉面積指數(shù)(Leaf Area Index,LAI)通過式(1)計(jì)算:
式中為折算系數(shù)(本文玉米取0.75,向日葵取0.65)[34-35];為種植密度,株/m2(本文玉米和向日葵分別為7.5和4株/m2);代表作物編號(hào),為對(duì)應(yīng)作物的葉片序號(hào),其中為第株作物的總?cè)~片數(shù);L為葉長(zhǎng),m;B為最大葉寬,m。
1.2.1 監(jiān)測(cè)數(shù)據(jù)特征提取
溫度數(shù)據(jù)甄別與區(qū)分的目標(biāo)是將熱紅外傳感器所測(cè)下墊面10個(gè)點(diǎn)位掃描溫度區(qū)分為作物冠層溫度(T)和土壤表面溫度(T)。玉米和向日葵生育前期LAI和較低,植被處于相對(duì)稀疏的狀態(tài),監(jiān)測(cè)系統(tǒng)每次掃描下墊面所測(cè)數(shù)據(jù)可能是T和T的混合結(jié)果。以2021年7月17日布置在永濟(jì)試驗(yàn)站玉米和向日葵田間的CTMS-On line系統(tǒng)獲得的日內(nèi)典型時(shí)段的溫度為例(圖 2),日內(nèi)每個(gè)時(shí)段(10:00—18:00)所掃描的10個(gè)溫度數(shù)據(jù)之間的標(biāo)準(zhǔn)方差大于0.1。7月17日玉米處于生育中期而向日葵開始進(jìn)入快速發(fā)育期,2種作物株高和葉面積指數(shù)差異較大,因而2種作物監(jiān)測(cè)結(jié)果差異較大。這種差異也間接地反映出2種作物田間墑情、生長(zhǎng)環(huán)境條件的不同,因而甄別和區(qū)分算法要因作物不同而調(diào)整。
圖2 玉米和向日葵田間CTMS-On line系統(tǒng)掃描的10個(gè)點(diǎn)位農(nóng)田下墊面溫度數(shù)據(jù)(2021年7月17日)
一般情況下,作物出苗后冠層高度隨日序數(shù)增加呈現(xiàn)“緩慢增長(zhǎng)—快速增長(zhǎng)—緩慢增長(zhǎng)”的趨勢(shì),LAI則呈現(xiàn)“緩慢增長(zhǎng)—快速增長(zhǎng)—緩慢增長(zhǎng)—快速減少”的趨勢(shì)。為了獲得準(zhǔn)確的作物生長(zhǎng)特征以構(gòu)建數(shù)據(jù)甄別算法,充分考慮作物冠層高度和葉面積指數(shù)在生育期內(nèi)主要特征,本文以日序數(shù)為自變量,利用2種Logistic生長(zhǎng)模型分別模擬每日和LAI的變化,其表達(dá)式為[36-38]
式中1和2分別代表作物生育期內(nèi)每日的和LAI;為日序數(shù),本文取年內(nèi)自然日;和分別為一定環(huán)境條件下的作物最大(m)和LAI;、、、、為對(duì)應(yīng)的Logistic模型待定系數(shù)。
1.2.2 Logistic模型擬合
圖3是根據(jù)2021年作物生長(zhǎng)指標(biāo)實(shí)際測(cè)量結(jié)果擬合的基于Logistic生長(zhǎng)模型的和LAI變化曲線。根據(jù)作物生長(zhǎng)趨勢(shì),其中關(guān)鍵變化轉(zhuǎn)折點(diǎn)記為M1(快速發(fā)育期起點(diǎn))、M2(快速發(fā)育期中點(diǎn))、M3(平穩(wěn)生長(zhǎng)期起點(diǎn))和M4(生育期內(nèi)葉面積指數(shù)最大點(diǎn))??梢娪衩缀拖蛉湛谧魑锷捌冢∕1之前)的LAI及相對(duì)較低,隨后生育中期(M1和M3之間)葉面積迅速發(fā)展LAI增大,而在作物生育后期(M4之后)葉片逐漸凋萎,葉面積逐漸減少。
為確定和LAI變化曲線的拐點(diǎn)(為簡(jiǎn)化算法,假定和LAI的關(guān)鍵轉(zhuǎn)折點(diǎn)對(duì)應(yīng)日序數(shù)一致),對(duì)式(2)和式(3)進(jìn)行求導(dǎo),整理得:
注:M1,快速發(fā)育期起點(diǎn);M2,快速發(fā)育期中點(diǎn);M3,平穩(wěn)生長(zhǎng)期起點(diǎn);M4,生育期內(nèi)葉面積指數(shù)最大值點(diǎn)。例:M1:07-17(198)中,07-17為日期,198為日序數(shù)。其余同上。
此時(shí)式(4)和式(5)分別為和LAI的變化速率方程。令式(5)等于0(增長(zhǎng)速率為0),便可求得LAI的最大值(LAImax)),并記該點(diǎn)處的日序數(shù)為M4。二者對(duì)應(yīng)計(jì)算公式為
對(duì)式(4)進(jìn)行一階求導(dǎo),并令其等于0,即可求出最大增長(zhǎng)速率對(duì)應(yīng)的日序數(shù)M2。在M1之前和M3之后緩慢增長(zhǎng)趨于平緩,而在M1和M3之間,作物生長(zhǎng)迅速。為確定這兩個(gè)特征點(diǎn),可對(duì)式(2)進(jìn)行三階求導(dǎo),并令其等于0,便可求出生長(zhǎng)曲線上的兩個(gè)突變拐點(diǎn),即最大生長(zhǎng)階段對(duì)應(yīng)的日序數(shù)區(qū)間(M1,M3)??傻蒙鲜鲫P(guān)鍵點(diǎn)計(jì)算式如下:
從圖3可知,2021年永濟(jì)試驗(yàn)站實(shí)測(cè)數(shù)據(jù)擬合的Logistic模型關(guān)鍵參數(shù),對(duì)于玉米:=1.27×1010,=0.118,=4.50,=1.20×10-3,=-0.557,=63.4;對(duì)于向日葵:=4.77×1010,=0.119,=6.68,=1.54×10-3,=-0.747,=90.3。結(jié)合式(6)和式(7),可推求關(guān)鍵節(jié)點(diǎn)對(duì)應(yīng)的日序數(shù)分別為:玉米:M1=186(7月5日),M2=198(7月17日),M3=209(7月28日),M4=232(8月20日),LAImax=3.52;向日葵:M1=198(7月17日),M2=207(7月26日),M3=218(8月7日),M4=243(8月31日),LAImax=3.99。
1.2.3 溫度數(shù)據(jù)甄別算法與流程
將作物生長(zhǎng)指標(biāo)和LAI作為首要評(píng)判指標(biāo),以作物生長(zhǎng)關(guān)鍵節(jié)點(diǎn)的日序數(shù)M1,M2,M3和作物L(fēng)AI最大時(shí)的日序數(shù)M4為時(shí)間節(jié)點(diǎn)(對(duì)應(yīng)圖3中的關(guān)鍵生長(zhǎng)節(jié)點(diǎn)(星號(hào)*)),來(lái)判斷和構(gòu)建溫度數(shù)據(jù)甄別算法。本文選取標(biāo)準(zhǔn)方差(Standard Deviation,SD)和相對(duì)誤差(Relative Error,RE)進(jìn)行相關(guān)數(shù)據(jù)的統(tǒng)計(jì)和分析。
本文設(shè)計(jì)溫度甄別和區(qū)分的算法流程見圖4,其中是需要甄別處理溫度數(shù)據(jù)當(dāng)天的年內(nèi)日序數(shù)。數(shù)據(jù)處理流程如下:
1)若M3≤≤M4,作物和LAI較高,長(zhǎng)勢(shì)良好枝葉茂密,那么10個(gè)掃描點(diǎn)位溫度默認(rèn)為葉片溫度,并以其平均值作為T。當(dāng)≥M4時(shí),由圖3可知作物株高基本保持穩(wěn)定,LAI呈現(xiàn)逐漸下降趨勢(shì),但在作物收獲前LAI的值仍較高,說明此時(shí)植株枝葉仍較密,10個(gè)掃描點(diǎn)位溫度仍主要掃描為葉片溫度,同樣以其平均值作為T。
2)當(dāng)≤M1時(shí),作物和LAI均較低,此時(shí)作物枝葉較小處于苗期,監(jiān)測(cè)系統(tǒng)所掃描的10個(gè)溫度數(shù)據(jù)主要來(lái)源于農(nóng)田土壤表面,少量點(diǎn)位掃描在葉片上。本時(shí)期內(nèi)的溫度數(shù)據(jù)篩選與區(qū)分,做如下處理:若10個(gè)掃描溫度的標(biāo)準(zhǔn)方差SD≤0.1(數(shù)據(jù)離散程度較小,基本穩(wěn)定),則所有值的均值默認(rèn)為T,此時(shí)T的獲取需通過人工測(cè)量溫度確定T與T之間的比例關(guān)系,以近似反推T。在溫度數(shù)據(jù)穩(wěn)定可靠的前提下,若10個(gè)掃描溫度的標(biāo)準(zhǔn)方差SD>0.1(數(shù)據(jù)離散程度稍大,具有一定程度區(qū)分性),可將溫度中最小值(min)臨近的幾個(gè)數(shù)值點(diǎn)均值作為T,其余的溫度數(shù)值點(diǎn)的均值作為T。
注:M1~M4分別為作物生長(zhǎng)關(guān)鍵節(jié)點(diǎn)的日序數(shù);SD為標(biāo)準(zhǔn)方差。
3)當(dāng)M1<
為整體數(shù)據(jù)集矩陣×矩陣,即由原始數(shù)據(jù)所組成,對(duì)應(yīng)每一時(shí)刻點(diǎn)所得監(jiān)測(cè)數(shù)據(jù)的均值所構(gòu)成的1×矩陣。指定計(jì)算主要包括bsxfun函數(shù)中的函數(shù)操作@ge和@le,@ge為大于或等于閾值,@le為小于或等于閾值。本文以CTMS-On line系統(tǒng)監(jiān)測(cè)的10個(gè)掃描溫度的平均值作為閾值,對(duì)原始溫度數(shù)據(jù)矩陣進(jìn)行區(qū)分,其他函數(shù)操作詳見https://ww2.mathworks.cn/help/matlab/ref/bsxfun. html?s_tid=srchtitle_bsxfun_1。
為確保區(qū)分后的T和T數(shù)據(jù)矩陣真實(shí)可靠,應(yīng)進(jìn)一步優(yōu)化校正。利用人工測(cè)量溫度數(shù)據(jù)與經(jīng)過甄別算法區(qū)分所得T和T數(shù)據(jù)之間的線性關(guān)系,確定作物T和T數(shù)據(jù)矩陣的校正因子(本文數(shù)據(jù)處理后可得冠層溫度校正因子玉米為0.9、向日葵為0.7,土壤表面溫度校正因子分別為1.1和1.2)。最后將區(qū)分后的T和T數(shù)據(jù)分別乘以相對(duì)應(yīng)的校正因子,即可得到最終甄別篩選的T和T數(shù)據(jù)。
選取永濟(jì)試驗(yàn)站玉米和向日葵生育期4個(gè)關(guān)鍵節(jié)點(diǎn)(M1,M2,M3和M4)的溫度數(shù)據(jù)進(jìn)行驗(yàn)證;通過分析溫度數(shù)據(jù)區(qū)分結(jié)果與人工測(cè)量數(shù)據(jù)的匹配情況,探究甄別算法的可行性和有效性。
2.1.1 玉米地
人工測(cè)量溫度在10:00—18:00內(nèi)間隔2 h進(jìn)行,觀測(cè)日內(nèi)可得目標(biāo)田塊土壤表面溫度T和作物冠層溫度T各5組數(shù)據(jù)。圖5是玉米和向日葵生育期內(nèi)4個(gè)關(guān)鍵時(shí)間節(jié)點(diǎn)人工測(cè)量溫度與數(shù)據(jù)甄別優(yōu)化后所得T和T的結(jié)果對(duì)比。其中,圖5為玉米地對(duì)比結(jié)果;由于玉米地在7月5日(M1)處于灌溉狀態(tài),此時(shí)T和T獲取精度難以保證,因而選擇田塊表面干燥后的7月8日(M1′)相對(duì)應(yīng)溫度數(shù)據(jù)進(jìn)行驗(yàn)證。
注:在關(guān)鍵節(jié)點(diǎn)(M)無(wú)實(shí)際田間溫度時(shí)(灌溉或降雨),用臨近節(jié)點(diǎn)(M1′)處驗(yàn)證。
從圖5可見,玉米的T和T均在14:00達(dá)到峰值,呈現(xiàn)明顯的單峰型變化特征。表面溫度監(jiān)測(cè)數(shù)據(jù)包含了T和T混合數(shù)據(jù),經(jīng)算法優(yōu)化后甄別區(qū)分出T和T,與人工測(cè)量T和T之間的變化趨勢(shì)保持一致,且大小較為接近;與沒有甄別的混合數(shù)據(jù)相比,匹配度明顯提高。4個(gè)生長(zhǎng)節(jié)點(diǎn)中快速發(fā)育期的M2處表現(xiàn)出更好的匹配結(jié)果。生育中后期的M4處玉米枝葉茂密長(zhǎng)勢(shì)良好,監(jiān)測(cè)系統(tǒng)所掃描的10個(gè)點(diǎn)位在作物葉片上;此時(shí)根據(jù)本文設(shè)計(jì)的甄別算法判定其全部為作物冠層溫度(等于甄別前監(jiān)測(cè)溫度),所平均得到的T數(shù)值與人工測(cè)量T的保持很高的匹配度(圖5)。與冠層溫度甄別對(duì)比結(jié)果相比,田間土壤表面溫度T在中午12:00—14:00差異稍大,甄別后所得的T均小于人工測(cè)量T;與冠層溫度甄別結(jié)果一致,甄別后T亦在M2處與人工測(cè)量T的匹配度最佳。
以人工測(cè)量溫度數(shù)據(jù)為參照量,分別計(jì)算上述時(shí)段溫度數(shù)據(jù)甄別優(yōu)化前后與其之間的相對(duì)誤差,同時(shí)計(jì)算RE優(yōu)化幅度,即甄別前RE絕對(duì)值與甄別后RE絕對(duì)值之間的差值;其值若為正值,數(shù)值越大表明精度提升幅度越大,若為負(fù)值則表明沒有提升精度,甄別效果不明顯。2021年永濟(jì)試驗(yàn)站玉米和向日葵地計(jì)算結(jié)果如圖6所示。由圖 6a可知,各節(jié)點(diǎn)處日內(nèi)甄別優(yōu)化后T與人工測(cè)量T的RE絕對(duì)值明顯低于甄別優(yōu)化前;M1′(7月8日)14:00和16:00處作物冠層溫度RE值,從甄別優(yōu)化前的19.0%和21.8%下降為甄別優(yōu)化后的-2.9%和1.8%,優(yōu)化幅度分別為16.0和19.9個(gè)百分點(diǎn)。其中M2和M3處各時(shí)刻甄別優(yōu)化后RE均相對(duì)較小,與前述分析結(jié)果一致。且各時(shí)間段內(nèi)計(jì)算的RE優(yōu)化幅度均為正值,表明甄別優(yōu)化后對(duì)溫度數(shù)據(jù)的精度有所提升,且在作物前中期優(yōu)化幅度最大(平均RE優(yōu)化幅度為12.2個(gè)百分點(diǎn))。與T表現(xiàn)類似,甄別優(yōu)化后T與人工測(cè)量T的平均RE絕對(duì)值從15%降為5%;土壤表面溫度與人工測(cè)量T之間的RE優(yōu)化幅度在7個(gè)百分點(diǎn)左右,溫度數(shù)據(jù)精度提升較高(圖6b)。
圖6 永濟(jì)試驗(yàn)站玉米和向日葵地表面溫度甄別優(yōu)化前后溫度結(jié)果與人工測(cè)量溫度之間RE變化
從總的RE值表現(xiàn)來(lái)看,7月8日M1′處18:00處甄別優(yōu)化后T和T的RE絕對(duì)值明顯高于其他時(shí)刻,其值分別為15.9%和12.5%,優(yōu)化幅度分別為14.1和-10.2個(gè)百分點(diǎn)。上述結(jié)果產(chǎn)生原因可能是由于本區(qū)域農(nóng)田常伴有冷氣流現(xiàn)象(多云或刮風(fēng)),導(dǎo)致此時(shí)田間溫度瞬時(shí)變化較大;同時(shí)人工測(cè)量與系統(tǒng)監(jiān)測(cè)之間的等時(shí)性也可能有一定的偏差。除此之外,總體上T和T經(jīng)甄別優(yōu)化后,RE優(yōu)化幅度大多為正值,即玉米地的T和T的平均優(yōu)化幅度分別為12.2和7.0個(gè)百分點(diǎn),其精度均得以有效提升高,表明甄別算法的可靠性。
2.1.2 向日葵地
為確定本文溫度甄別方法在向日葵地的應(yīng)用效果,對(duì)向日葵的4個(gè)典型生長(zhǎng)節(jié)點(diǎn)(M1,M2,M3和M4)處所獲得的甄別前后的T和T與人工測(cè)量溫度數(shù)據(jù)進(jìn)行對(duì)比驗(yàn)證。如圖5向日葵作物所示,在10:00—18:00時(shí)間段內(nèi)向日葵地T和T呈現(xiàn)單值波峰現(xiàn)象,且甄別優(yōu)化后T和T與人工測(cè)量T和T的變化趨勢(shì)基本一致,典型時(shí)刻內(nèi)數(shù)據(jù)結(jié)果相差不大。與玉米地類似,向日葵在快速發(fā)育期節(jié)點(diǎn)M2(7月26日)處甄別后的T和T與人工測(cè)量數(shù)值的匹配程度最優(yōu),此時(shí)作物長(zhǎng)勢(shì)良好,農(nóng)田下墊面溫度數(shù)據(jù)被有效的區(qū)分為T和T。而當(dāng)作物處于生長(zhǎng)中后期,作物枝葉茂密長(zhǎng)勢(shì)良好有效遮蔽土壤表面,甄別算法確定此時(shí)監(jiān)測(cè)溫度都是冠層溫度T,與人工測(cè)量T匹配結(jié)果良好(圖5h)。
從向日葵田塊下墊面溫度數(shù)據(jù)甄別優(yōu)化前后與人工測(cè)量溫度之間的相對(duì)誤差RE變化(圖6c和圖6d)看,M1(7月17日)和M2(7月26日)處,T的RE絕對(duì)值較甄別優(yōu)化前降低了30%,且RE優(yōu)化幅度也相對(duì)較高均大于20個(gè)百分點(diǎn),整體的平均優(yōu)化幅度可達(dá)32.3個(gè)百分點(diǎn)。與此同時(shí),甄別優(yōu)化后T的RE值穩(wěn)定在-5%~5%,除M1節(jié)點(diǎn)16:00和18:00時(shí)間段內(nèi)RE優(yōu)化幅度較低,整體的平均優(yōu)化幅度仍有12.0個(gè)百分點(diǎn)。在向日葵生育中后期作物長(zhǎng)勢(shì)茂密,監(jiān)測(cè)數(shù)據(jù)甄別的T和T與人工測(cè)量數(shù)值之間的匹配效果雖略弱于前,但它們之間的RE絕對(duì)值仍可保持在10%以內(nèi)(M3(8月7日)),符合預(yù)期的溫度數(shù)據(jù)甄別精度。
為進(jìn)一步查勘農(nóng)田表面溫度數(shù)據(jù)甄別算法的效果,以內(nèi)蒙古河套灌區(qū)解放閘灌域和吉林省長(zhǎng)春試驗(yàn)站的CTMS-Online系統(tǒng)監(jiān)測(cè)的農(nóng)田下墊面溫度數(shù)據(jù)為例進(jìn)行校核和驗(yàn)證。其中長(zhǎng)春試驗(yàn)站為2018年和2019年玉米地監(jiān)測(cè)數(shù)據(jù),解放閘灌域是2016年玉米和向日葵兩種作物的數(shù)據(jù)。根據(jù)當(dāng)年的農(nóng)田實(shí)測(cè)LAI和數(shù)據(jù),兩個(gè)地區(qū)玉米和向日葵所擬合的Logistic生長(zhǎng)模型模擬曲線及關(guān)鍵生長(zhǎng)節(jié)點(diǎn)如圖7所示。
圖7 解放閘灌域和長(zhǎng)春試驗(yàn)站基于玉米和向日葵h和LAI的Logistic生長(zhǎng)模型擬合結(jié)果
2.2.1 長(zhǎng)春玉米地校核與驗(yàn)證
從圖7a和圖7b中可知,長(zhǎng)春試驗(yàn)站玉米地2018年與2019年利用Logistic生長(zhǎng)模型確定的M1、M2、M3和M4的關(guān)鍵生長(zhǎng)節(jié)點(diǎn)(日序數(shù))差別不大,兩年日期基本相臨。因慮及同一地區(qū)同類作物,這里選擇2018年模型擬合對(duì)應(yīng)的關(guān)鍵節(jié)點(diǎn)日期應(yīng)用于溫度數(shù)據(jù)甄別篩選算法的校核與驗(yàn)證。由于2018年節(jié)點(diǎn)M2(7月9日)連續(xù)多天降雨,為保證數(shù)據(jù)驗(yàn)證的可靠性,以M2′處(7月12日)替代;其余關(guān)鍵節(jié)點(diǎn)M1(6月22日)、M3(7月24日)和M4(8月16日)如圖7a所示。這里統(tǒng)計(jì)了甄別優(yōu)化前后T和T與人工測(cè)量溫度之間的相對(duì)誤差RE值,并計(jì)算了RE優(yōu)化幅度,如表1所示。
從表1可知,2018年冠層溫度T在M1處RE從19.6%降到7.7%,有11.9個(gè)百分點(diǎn)的優(yōu)化幅度;M2′和M3處分別從9.4%降至1.5%和從12.8%降至2.3%,優(yōu)化幅度分別為7.9和10.5個(gè)百分點(diǎn),精度提升效果較為明顯;2019年優(yōu)化幅度則分別為10.3(M1)、2.7(M2)和1.4(M3)個(gè)百分點(diǎn)。2018年在M2′處和2019年M2處甄別優(yōu)化后T與人工測(cè)量T之間的RE絕對(duì)值<5%,與永濟(jì)試驗(yàn)站甄別結(jié)果最佳時(shí)段一致。M4節(jié)點(diǎn)玉米處于生長(zhǎng)中后期,數(shù)據(jù)甄別程序判斷下墊面掃描溫度均為冠層溫度;除2019年M4處18:00時(shí)刻與人工測(cè)量T的RE絕對(duì)值稍大(9.7%),其他時(shí)刻內(nèi)均保持在5%以內(nèi)。土壤表面溫度T的甄別驗(yàn)證結(jié)果則表現(xiàn)為,RE絕對(duì)值大致從10%左右縮減到5%以內(nèi),能夠與人工測(cè)量T相貼近;各階段整體的的優(yōu)化幅度大約為5個(gè)百分點(diǎn),且甄別效果最佳階段亦在M2階段。其中因生育早期甄別前T與人工測(cè)量T較為接近,通過算法確定出的T精度提升不明顯,如2018年M1處??傮w而言,2018 —2019年長(zhǎng)春玉米T和T平均優(yōu)化幅度分別為7.5%和5.5%,本文的甄別算法滿足長(zhǎng)春試驗(yàn)站玉米地T和T的甄別區(qū)分精度要求。
2.2.2 解放閘灌域玉米和向日葵校核與驗(yàn)證
解放閘灌域和永濟(jì)灌域同屬于河套灌區(qū),主要農(nóng)作物為玉米和向日葵。這里利用2016年解放閘灌域所監(jiān)測(cè)的兩種作物生育期下墊面溫度數(shù)據(jù)進(jìn)行甄別處理,并將甄別優(yōu)化前后的T和T與同時(shí)段的人工測(cè)量溫度進(jìn)行對(duì)比,以查核甄別算法在本區(qū)域的應(yīng)用效果。
從圖3a和圖7c可知,玉米生長(zhǎng)指標(biāo)數(shù)值在解放閘灌域和永濟(jì)灌域兩地比較相近(永濟(jì):LAImax=3.52,max=2.72 m;解放閘:LAImax=3.72,max=2.61 m),因此甄別算法中采用同一校正因子(=0.9,=1.1)計(jì)算玉米地甄別優(yōu)化前后T和T與人工測(cè)量值之間的RE值及其優(yōu)化幅度,結(jié)果如圖8a和圖8b所示。由圖 8a可見,玉米生育中后期(M3,M4)長(zhǎng)勢(shì)穩(wěn)定,甄別優(yōu)化后T與人工測(cè)量T之間的RE絕對(duì)值低于5%,兩者匹配度較高;生育前中期(M1,M2)甄別優(yōu)化后的T與人工測(cè)量T之間的RE值明顯低于甄別優(yōu)化前的結(jié)果,其絕對(duì)值在10%以內(nèi)。除了M1和M2節(jié)點(diǎn)處10:00時(shí)段優(yōu)化幅度為負(fù)值,其他時(shí)間段內(nèi)均為正值,這表明甄別優(yōu)化后的T能與人工測(cè)量T較好的契合,其精度得以保證。圖 8b中甄別優(yōu)化后的T與人工測(cè)量T之間的RE絕對(duì)值在10%以內(nèi),且除了M1和M2節(jié)點(diǎn)處16:00時(shí)段優(yōu)化幅度為負(fù)值,其他時(shí)間段內(nèi)均為正值;甄別優(yōu)化后T和T的整體平均優(yōu)化幅度分別為4.9和7.5個(gè)百分點(diǎn),表明該甄別方法甄別精度較為理想。
表1 長(zhǎng)春試驗(yàn)站玉米地監(jiān)測(cè)數(shù)據(jù)甄別優(yōu)化前后的溫度與人工測(cè)量之間相對(duì)誤差變化
解放閘灌域向日葵溫度數(shù)據(jù)甄別,首先嘗試采用與永濟(jì)灌域同一校正因子(=0.7,=1.2),甄別優(yōu)化前后T和T與人工測(cè)量值之間的相對(duì)誤差RE值及其優(yōu)化幅度結(jié)果如圖8c和圖8d所示。從圖中可見,除了M1和M2處T甄別優(yōu)化后的RE絕對(duì)值較小,其余節(jié)點(diǎn)和日內(nèi)時(shí)段均超過了10%,且T和T的RE優(yōu)化幅度多數(shù)小于0。數(shù)值不符合預(yù)期設(shè)想,沒有達(dá)到精度提升的效果。這種情況可能是由于解放閘灌域相較于永濟(jì)灌域的向日葵LAI和株高更大(圖3b和圖7d,永濟(jì):LAImax=3.99,max=1.77 m;解放閘:LAImax=4.56,max=2.21 m),CTMS-On line系統(tǒng)熱紅外傳感器運(yùn)行時(shí)連續(xù)掃描下墊面,在作物覆蓋度較高時(shí)掃描到葉片上的概率較大(旋轉(zhuǎn)臂均勻旋轉(zhuǎn)一周獲取10個(gè)溫度數(shù)據(jù)),掃描到土壤表面的概率也就越小,采用相同校正因子會(huì)對(duì)結(jié)果產(chǎn)生一定誤差。因此,針對(duì)永濟(jì)灌域向日葵冠層溫度校正因子(=0.7)和土壤表面溫度校正因子(=1.2),適當(dāng)調(diào)整并設(shè)定解放閘灌域校正因子(=0.9,=1.1)。利用此設(shè)定的校正因子進(jìn)行甄別優(yōu)化區(qū)分,結(jié)果如圖8e和圖8f所示??梢姅?shù)據(jù)甄別優(yōu)化后RE值的紅色柱大幅縮短,其絕對(duì)值基本在5%以內(nèi),整體上RE優(yōu)化幅度為正值,T和T的平均優(yōu)化幅度分別為8.8和6.3個(gè)百分點(diǎn),較校正前均得到大幅提升。與圖8c和圖8d相比,此時(shí)甄別優(yōu)化后的溫度數(shù)據(jù)與人工測(cè)量溫度數(shù)據(jù)吻合度較高,使得T和T的整體精度均得以大幅度提升。由此可知,同一作物的不同生長(zhǎng)情況下,數(shù)據(jù)甄別算法應(yīng)根據(jù)相應(yīng)的LAImax確定適宜的校正因子數(shù)值。
圖8 解放閘灌域玉米和向日葵地甄別優(yōu)化后的溫度與人工測(cè)量之間相對(duì)誤差
農(nóng)田下墊面溫度自動(dòng)監(jiān)測(cè)系統(tǒng)包含了土壤表面溫度T和作物冠層溫度T的混合數(shù)據(jù),及時(shí)和合理的利用甄別區(qū)分算法,對(duì)于監(jiān)測(cè)系統(tǒng)在處理海量數(shù)據(jù)時(shí)至關(guān)重要。本文考慮了作物的生長(zhǎng)狀態(tài)以及關(guān)鍵節(jié)點(diǎn)來(lái)建立甄別算法,取得了較好效果。
根據(jù)種植密度和植被覆蓋度的不同,作物可大致分為稀疏植被和密集植被[39-40]。除本文所分析的向日葵和玉米外,番茄、大豆以及辣椒等在生長(zhǎng)前中期屬于稀疏植被范疇內(nèi)的對(duì)象,亦可考慮采用本文溫度甄別篩選方法對(duì)利用熱紅外傳感器掃描獲得的溫度數(shù)據(jù)進(jìn)行有效甄別。從解放閘灌域向日葵溫度數(shù)據(jù)甄別區(qū)分過程可知,同一種作物在不同種植密度及其不同的LAI變化情況下,要適當(dāng)調(diào)整校正因子,實(shí)現(xiàn)不同種植情況下的有效甄別,且獲取的T和T均符合精度要求。參照永濟(jì)試驗(yàn)站與解放閘灌域的實(shí)測(cè)數(shù)據(jù),本文以LAImax等于4為基礎(chǔ)每向上增加1,向日葵冠層溫度校正因子調(diào)高0.35,土壤表面溫度校正因子調(diào)低0.18,精度提升明顯。
本研究基于3個(gè)地區(qū)(永濟(jì)、長(zhǎng)春、解放閘)進(jìn)行了相關(guān)研究,計(jì)算結(jié)果表現(xiàn)了良好的一致性,最終可以根據(jù)作物種植密度及LAI合理確定校正因子數(shù)值。在作物不同生育期,溫度的變化是具有差異性的,因而實(shí)際的校正因子可能是一個(gè)動(dòng)態(tài)變化值。這也是后續(xù)需要進(jìn)一步研究的地方,初步設(shè)想將其與田間氣溫的動(dòng)態(tài)變化相結(jié)合,以獲取更為準(zhǔn)確的校正因子。在沒有實(shí)測(cè)數(shù)據(jù)的情況下,推薦本文數(shù)值處理方法作為參考。隨后將進(jìn)一步對(duì)星載的大范圍數(shù)據(jù)進(jìn)行處理和計(jì)算分析,對(duì)甄別算法進(jìn)行校核驗(yàn)證與完善。
在農(nóng)田信息監(jiān)測(cè)方面,無(wú)人機(jī)遙感具有高頻、迅捷、空間分辨率高、時(shí)效性強(qiáng)等特點(diǎn)。且在精準(zhǔn)農(nóng)業(yè)領(lǐng)域中,??捎糜诒O(jiān)測(cè)田塊尺度上的作物生長(zhǎng)指標(biāo)(植被覆蓋度、葉面積指數(shù)、株高等),并可有效預(yù)測(cè)作物產(chǎn)量[41-42]。無(wú)人機(jī)遙感利用熱紅外傳感器獲取2.5~14m波段連續(xù)和非連續(xù)的數(shù)據(jù),在植被覆蓋度較高的地區(qū),無(wú)人機(jī)遙感獲取的地表溫度被默認(rèn)為作物冠層溫度,可結(jié)合作物水分脅迫指數(shù)(Crop Water Stress Index,CWSI)量化作物含水率與冠層溫度的關(guān)系,進(jìn)而反映農(nóng)田水分狀況[42-45]。在植被覆蓋度中等的地區(qū),無(wú)人機(jī)遙感獲取的地表溫度是冠層與土壤溫度的混合體,且土壤表面溫度對(duì)監(jiān)測(cè)作物冠層溫度有著不可忽視的影響,對(duì)于如何有效精量甄別無(wú)人機(jī)遙感獲取的下墊面溫度的相關(guān)研究較少。本文所設(shè)計(jì)的考慮作物生長(zhǎng)狀態(tài)的農(nóng)田表面溫度數(shù)據(jù)甄別方法,可為無(wú)人機(jī)多種植被類型的遙感溫度數(shù)據(jù)精量甄別提供幫助;同理,對(duì)衛(wèi)星遙感數(shù)據(jù)進(jìn)行合理甄別與區(qū)分,為估算大尺度灌域的蒸散發(fā)提供精細(xì)的溫度數(shù)據(jù)來(lái)源。
通過考慮作物生長(zhǎng)狀態(tài)指標(biāo)和關(guān)鍵節(jié)點(diǎn),利用作物生長(zhǎng)模型,設(shè)計(jì)了農(nóng)田表面溫度數(shù)據(jù)進(jìn)行甄別與區(qū)分算法,并利用田間監(jiān)測(cè)數(shù)據(jù)進(jìn)行校核和驗(yàn)證,得到如下結(jié)論:
1)作物生長(zhǎng)指標(biāo)葉面積指數(shù)LAI和冠層高度可作為稀疏植被溫度數(shù)據(jù)甄別有效區(qū)間的判定指標(biāo);結(jié)合Logistic作物生長(zhǎng)模型,確定作物生育期關(guān)鍵節(jié)點(diǎn)來(lái)設(shè)計(jì)合理數(shù)據(jù)甄別算法。
2)通過與人工測(cè)量值對(duì)比,甄別方法對(duì)冠層溫度(T)和)土壤表面溫度(T)的獲取精度得以有效提升。以二者之間相對(duì)誤差(RE)變化為例,永濟(jì)試驗(yàn)站玉米T和T優(yōu)化幅度為12.2和7.0個(gè)百分點(diǎn),向日葵優(yōu)化幅度為32.3和12.0個(gè)百分點(diǎn);兩年長(zhǎng)春玉米T和T的平均優(yōu)化幅度分別為7.5和5.5個(gè)百分點(diǎn);解放閘灌域玉米T和T優(yōu)化幅度分別為4.9和7.5個(gè)百分點(diǎn),經(jīng)調(diào)整甄別算法中校正因子后向日葵T和T優(yōu)化幅度為8.8和6.3個(gè)百分點(diǎn)。
3)甄別算法中校正因子數(shù)值需根據(jù)作物種植密度及LAI確定;玉米校正因子選擇作物冠層溫度校正因子0.9,土壤表面溫度校正因子1.1;向日葵校正因子以葉面積指數(shù)最大值4為基礎(chǔ),選取冠層溫度校正因子0.7,土壤表面溫度校正因子1.2;在不同地區(qū)應(yīng)用時(shí),向日葵葉面積指數(shù)最大值每增加1,推薦冠層溫度校正因子調(diào)高0.35,土壤表面溫度校正因子調(diào)低0.18。
[1] 陳媛媛,游炯,幸澤峰,等. 世界主要國(guó)家精準(zhǔn)農(nóng)業(yè)發(fā)展概況及對(duì)中國(guó)的發(fā)展建議[J]. 農(nóng)業(yè)工程學(xué)報(bào),2021,37(11):315-324.
Chen Yuanyuan, You Jiong, Xing Zefeng, et al. Review of precision agriculture development situations in the main countries in the world and suggestions for China[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2021, 37(11): 315-324. (in Chinese with English abstract)
[2] 何東健,楊成海,楊青,等. 面向精準(zhǔn)農(nóng)業(yè)的農(nóng)田土壤成分實(shí)時(shí)測(cè)定研究進(jìn)展[J]. 農(nóng)業(yè)工程學(xué)報(bào),2012,28(7):78-85.
He Dongjian, Yang Chenghai, Yang Qing, et al. Research progress of real-time measurement of soil attributes for precision agriculture[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2012, 28(7): 78-85. (in Chinese with English abstract)
[3] Bwambale E, Abagale F K, Anornu G K. Smart irrigation monitoring and control strategies for improving water use efficiency in precision agriculture: A review[J]. Agricultural Water Management, 2022, 260: 107324.
[4] 魯旭濤,張麗娜,劉昊,等. 智慧農(nóng)業(yè)水田作物網(wǎng)絡(luò)化精準(zhǔn)灌溉系統(tǒng)設(shè)計(jì)[J]. 農(nóng)業(yè)工程學(xué)報(bào),2021,37(17):71-81.
Lu Xutao, Zhang Lina, Liu Hao, et al. Design of the networked precision irrigation system for paddy field crops in intelligent agriculture[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2021, 37(17): 71-81. (in Chinese with English abstract)
[5] 蔡甲冰,張寶忠,劉鈺. 精量灌溉決策與控制技術(shù)[J]. 中國(guó)水利,2016(9):66-67.
Cai Jiabing, Zhang Baozhong, Liu Yu. Precision irrigation decision and control technology[J]. China Water Resources, 2016(9): 66-67. (in Chinese with English abstract)
[6] 李淑華. 農(nóng)田土壤墑情監(jiān)測(cè)與智能灌溉云服務(wù)平臺(tái)構(gòu)建關(guān)鍵技術(shù)研究[D]. 北京:中國(guó)農(nóng)業(yè)科學(xué)院,2016.
Li Shuhua. Research on Key Technologies of Cloud Service Platform for Farmland Foil Moisture Monitoring and Intelligent Irrigation[D]. Beijing: Chinese Academy of Agricultural Sciences, 2016. (in Chinese with English abstract)
[7] 賈敬敦,魯相潔,黃峰,等. 遠(yuǎn)程控制與無(wú)線通信技術(shù)在農(nóng)業(yè)中的應(yīng)用分析與展望[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào),2021,52(S1):351-359.
Jia Jingdun, Lu Xiangjie, Huang Feng, et al. Review of remote-control technology and wireless communication technology in agricultural applications[J]. Transactions of the Chinese Society for Agricultural Machinery, 2021, 52(S1): 351-359. (in Chinese with English abstract)
[8] 喬曉軍,張馨,王成,等. 無(wú)線傳感器網(wǎng)絡(luò)在農(nóng)業(yè)中的應(yīng)用[J]. 農(nóng)業(yè)工程學(xué)報(bào),2005,21(增刊2):232-234.
Qiao Xiaojun, Zhang Xin, Wang Cheng, et al. Application of the wireless sensor networks in agriculture[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2005, 21(Supp.2): 232-234. (in Chinese with English abstract)
[9] Hamami L, Nassereddine B. Application of wireless sensor networks in the field of irrigation: A review[J]. Computers and Electronics in Agriculture, 2020, 179: 105782.
[10] 武永峰,宮志宏,劉布春,等. 基于遠(yuǎn)程監(jiān)控的農(nóng)業(yè)氣象自動(dòng)采集系統(tǒng)設(shè)計(jì)[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào),2010,41(10):174-179.
Wu Yongfeng, Gong Zhihong, Liu Buchun, et al. Automatic collecting system of agro-meteorology information based on remote monitoring technique[J]. Transactions of the Chinese Society for Agricultural Machinery, 2010, 41(10): 174-179. (in Chinese with English abstract)
[11] 林蘭芬,王瑞松,于鵬華. 基于GIS的農(nóng)田小氣候環(huán)境可視監(jiān)測(cè)系統(tǒng)[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào),2015,46(3):254-260.
Lin Lanfen, Wang Ruisong, Yu Penghua. GIS based visual monitoring system of farmland microclimate environment[J]. Transactions of the Chinese Society for Agricultural Machinery, 2015, 46(3): 254-260. (in Chinese with English abstract)
[12] 盧健,陳桂芬,孟穎,等. 基于三維GIS的農(nóng)田環(huán)境模擬監(jiān)測(cè)的研究及系統(tǒng)實(shí)現(xiàn)[J]. 中國(guó)農(nóng)業(yè)科技導(dǎo)報(bào),2017,19(3):50-55.
Lu Jian, Chen Guifen, Meng Ying, et al. Simulation of farmland environmental monitoring based on 3D GIS and system implementation[J]. Journal of Agricultural Science and Technology, 2017, 19(3): 50-55. (in Chinese with English abstract)
[13] Sakamoto T, Gitelson A A, Nguy-Robertson A L, et al. An alternative method using digital cameras for continuous monitoring of crop status[J]. Agricultural and Forest Meteorology, 2011, 154-155: 113-126.
[14] Kim J, Ryu Y, Jiang C Y, et al. Continuous observation of vegetation canopy dynamics using an integrated low-cost, near-surface remote sensing system[J]. Agricultural and Forest Meteorology, 2019, 264: 164-177.
[15] Xiang H T, Tian L. An automated stand-alone in-field remote sensing system (SIRSS) for in-season crop monitoring[J]. Computers and Electronics in Agriculture, 2011, 78: 1-8.
[16] 李楠,劉成良,李彥明,等. 基于3S技術(shù)聯(lián)合的農(nóng)田墑情遠(yuǎn)程監(jiān)測(cè)系統(tǒng)開發(fā)[J]. 農(nóng)業(yè)工程學(xué)報(bào),2010,26(4):169-174.
Li Nan, Liu Chengliang, Li Yanming, et al. Development of remote monitoring system for soil moisture based on 3S technology alliance[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2010, 26(4): 169-174. (in Chinese with English abstract)
[17] 張?jiān)隽郑魰詰c. 基于ZigBee和ARM9的農(nóng)田墑情遠(yuǎn)程監(jiān)測(cè)系統(tǒng)[J]. 節(jié)水灌溉,2011(7):54-57.
Zhang Zenglin, Yu Xiaoqing. Remote monitoring system for soil moisture based on ZigBee and ARM9[J]. Water Saving Irrigation, 2011(7): 54-57. (in Chinese with English abstract)
[18] 蔡甲冰,許迪,司南,等. 基于冠層溫度和土壤墑情的實(shí)時(shí)監(jiān)測(cè)與灌溉決策系統(tǒng)[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào),2015,46(12):118-124.
Cai Jiabing, Xu Di, Si Nan, et al. Real-time monitoring system of crop canopy temperature and soil moisture for irrigation decision-making[J]. Transactions of the Chinese Society for Agricultural Machinery, 2015, 46(12): 118-124. (in Chinese with English abstract)
[19] Abioye E A, Abidin M S, Mahmud M S A, et al. A review on monitoring and advanced control strategies for precision irrigation[J]. Computers and Electronics in Agriculture, 2020, 173: 105441.
[20] 史長(zhǎng)麗,郭家選,梅旭榮,等. 夏玉米農(nóng)田表面溫度影響因素分析[J]. 中國(guó)農(nóng)業(yè)科學(xué),2006,39(1):48-56.
Shi Changli, Guo Jiaxuan, Mei Xurong, et al. Analysis of the factors influencing surface temperature in summer maize field[J]. Scientia Agricultura Sinica, 2006, 39(1): 48-56. (in Chinese with English abstract)
[21] Bense V F, Read T, Verhoef A. Using distributed temperature sensing to monitor field scale dynamics of ground surface temperature and related substrate heat flux[J]. Agricultural and Forest Meteorology, 2016, 220: 207-215.
[22] Jackson R D, Reginato R J, Idso S B. Wheat canopy temperature: A practical tool for evaluating water requirements[J]. Water Resources Research, 1977, 13(3): 651-656.
[23] 張宏鳴,王佳佳,韓文霆,等. 基于熱紅外遙感影像的作物冠層溫度提取[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào),2019,50(4):203-210.
Zhang Hongming, Wang Jiajia, Han Wenting, et al. Crop canopy temperature extraction based on thermal infrared remote sensing images[J]. Transactions of the Chinese Society for Agricultural Machinery, 2019, 50(4): 203-210. (in Chinese with English abstract)
[24] 張智韜,吳天奎,于廣多,等. 夏玉米冠層溫度變化的時(shí)滯效應(yīng)及其對(duì)土壤水分監(jiān)測(cè)的影響[J]. 農(nóng)業(yè)工程學(xué)報(bào),2022,38(1):117-124.
Zhang Zhitao, Wu Tiankui, Yu Guangduo, et al. Time delay effect of summer maize canopy temperature change and its influence on soil moisture content monitoring[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2022, 38(1): 117-124. (in Chinese with English abstract)
[25] 劉濤,仲曉春,孫成明. 作物溫度及其監(jiān)測(cè)技術(shù)研究進(jìn)展[J].中國(guó)農(nóng)業(yè)科技導(dǎo)報(bào),2017,19(12):59-66.
Liu Tao, Zhong Xiaochun, Sun Chengming. Review on crop temperature and its monitoring technologies[J]. Journal of Agricultural Science and Technology, 2017, 19(12): 59-66. (in Chinese with English abstract)
[26] Wei J X, Zhu W B. An operational parameterization scheme of surface temperature-vegetation index contextual model for large-scale temporally continuous evapotranspiration estimation: The case study of contiguous United States[J]. Journal of Hydrology, 2021, 602: 126805.
[27] Idso S B, Reginato R J, Jackson R D, et al. Measuring yield-reducing plant water potential depressions in wheat by infrared thermometry[J]. Irrigation Science, 1981, 2: 205-212.
[28] Miguel N, Borja M, Juan J P, et al. A new low-cost device based on thermal infrared sensors for olive tree canopy temperature measurement and water status monitoring[J]. Remote Sensing, 2020, 12(4): 723-723.
[29] 蔡甲冰,白亮亮,許迪,等. 基于地面紅外檢測(cè)系統(tǒng)驗(yàn)證的灌區(qū)地表溫度遙感反演[J]. 農(nóng)業(yè)工程學(xué)報(bào),2017,33(5):108-114.
Cai Jiabing, Bai Liangliang, Xu Di, et al. Remote sensing inversion of land surface temperature based on validation by observed infrared temperature in situ[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2017, 33(5): 108-114. (in Chinese with English abstract)
[30] Kool D, Agam N, Lazarovitch N, et al. A review of approaches for evapotranspiration partitioning[J]. Agricultural and Forest Meteorology, 2014, 184: 56-70.
[31] 張寶忠,許迪,劉鈺,等. 多尺度蒸散發(fā)估測(cè)與時(shí)空尺度拓展方法研究進(jìn)展[J]. 農(nóng)業(yè)工程學(xué)報(bào),2015,31(6):8-16.
Zhang Baozhong, Xu Di, Liu Yu, et al. Review of multi-scale evapotranspiration estimation and spatio-temporal scale expansion[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2015, 31(6): 8-16. (in Chinese with English abstract)
[32] 楊雨亭,尚松浩. 雙源蒸散發(fā)模型估算潛在蒸散發(fā)量的對(duì)比[J]. 農(nóng)業(yè)工程學(xué)報(bào),2012,28(24):85-91.
Yang Yuting, Shang Songhao. Comparison of dual-source evapotranspiration models in estimating potential evaporation and transpiration[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2012, 28(24): 85-91. (in Chinese with English abstract)
[33] Yang Y T, Shang S H. A hybrid dual‐source scheme and trapezoid framework–based evapotranspiration model (HTEM) using satellite images: Algorithm and model test[J]. Journal of Geophysical Research: Atmospheres, 2013, 118(5): 2284-2300.
[34] 史寶成. 作物缺水診斷指標(biāo)及灌溉控制指標(biāo)的研究[D]. 北京:中國(guó)水利水電科學(xué)研究院,2006.
Shi Baocheng. Experimental Study on Crop Water Stress Detecting and Irrigation Index[D]. Beijing: China Institute of Water Resources and Hydropower Research, 2006. (in Chinese with English abstract)
[35] 程煜. 有機(jī)無(wú)機(jī)肥配施對(duì)河套灌區(qū)重度鹽漬土理化性質(zhì)和向日葵生長(zhǎng)的影響[D]. 咸陽(yáng):西北農(nóng)林科技大學(xué),2022.
Cheng Yu. Effects of Combined Application of Organic and Inorganic Fertilizers on Physicochemical Properties of High Saline Soil and Growth of Sunflower in Hetao Irrigation District[D]. Xianyang: Northwest A&F University, 2022. (in Chinese with English abstract)
[36] Wu R L, Ma C X, Littell R C, et al. A logistic mixture model for characterizing genetic determinants causing differentiation in growth trajectories[J]. Genetical Research, 2002, 79(3): 235-245.
[37] 孫仕軍,姜浩,陳志君,等. 不同顏色地膜覆蓋下春玉米主要生長(zhǎng)性狀對(duì)耕層積溫的響應(yīng)[J]. 草業(yè)學(xué)報(bào),2019,28(2):61-72.
Sun Shijun, Jiang Hao, Chen Zhijun, et al. Effects of surface-layer accumulated temperature on major growth traits of spring maize when un-mulched or under clear or black plastic film mulches[J]. Acta Prataculturae Sinica, 2019, 28(2): 61-72. (in Chinese with English abstract)
[38] 蔡甲冰,常宏芳,陳鶴,等. 基于不同有效積溫的玉米干物質(zhì)累積量模擬[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào),2020,51(5):263-271.
Cai Jiabing, Chang Hongfang, Chen He, et al. Simulation of maize dry matter accumulation in normalized logistic model with different effective accumulated temperatures in field[J]. Transactions of the Chinese Society for Agricultural Machinery, 2020, 51(5): 263-271. (in Chinese with English abstract)
[39] 呂麗華,陶洪斌,夏來(lái)坤,等. 不同種植密度下的夏玉米冠層結(jié)構(gòu)及光合特性[J]. 作物學(xué)報(bào),2008,34(3):447-455.
Lü Lihua, Tao Hongbin, Xia Laikun, et al. Canopy structure and photosynthesis traits of summer maize under different planting densities[J]. Acta Agronomica Sinica, 2008, 34(3): 447-455. (in Chinese with English abstract)
[40] 郭柯,方精云,王國(guó)宏,等. 中國(guó)植被分類系統(tǒng)修訂方案[J]. 植物生態(tài)學(xué)報(bào),2020,44(2):111-127.
Guo Ke, Fang Jingyun, Wang Guohong, et al. A revised scheme of vegetation classification system of China[J]. Chinese Journal of Plant Ecology, 2020, 44(2): 111-127. (in Chinese with English abstract)
[41] 韓文霆,湯建棟,張立元,等. 基于無(wú)人機(jī)遙感的玉米水分利用效率與生物量監(jiān)測(cè)[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào),2021,52(5):129-141.
Han Wenting, Tang Jiandong, Zhang Liyuan, et al. Maize water use efficiency and biomass estimation based on unmanned aerial vehicle remote sensing[J]. Transactions of the Chinese Society for Agricultural Machinery, 2021, 52(5): 129-141. (in Chinese with English abstract)
[42] 紀(jì)景純,趙原,鄒曉娟,等. 無(wú)人機(jī)遙感在農(nóng)田信息監(jiān)測(cè)中的應(yīng)用進(jìn)展[J]. 土壤學(xué)報(bào),2019, 56(4):773-784.
Ji Jingchun, Zhao Yuan, Zou Xiaojuan, et al. Advancement in application of UAV remote sensing to monitoring of farmlands[J]. Acta Pedologica Sinica, 2019, 56(4): 773-784. (in Chinese with English abstract)
[43] Jackson R D, Idso S B, Reginato R J, et al. Canopy temperature as a crop water stress indicator[J]. Water Resources Research, 1981, 17(4): 1133-1138.
[44] 張智韜,于廣多,吳天奎,等. 基于無(wú)人機(jī)遙感影像的玉米冠層溫度提取及作物水分脅迫監(jiān)測(cè)[J]. 農(nóng)業(yè)工程學(xué)報(bào),2021,37(23):82-89.
Zhang Zhitao, Yu Guangduo, Wu Tiankui, et al. Temperature extraction of maize canopy and crop water stress monitoring based on UAV remote sensing images[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2021, 37(23): 82-89. (in Chinese with English abstract)
[45] 李志軍,于廣多,劉奇,等. 環(huán)境因子與玉米生長(zhǎng)對(duì)地表溫度監(jiān)測(cè)土壤水分的影響[J]. 農(nóng)業(yè)工程學(xué)報(bào),2022,38(20):77-85.
Li Zhijun, Yu Guangduo, Liu Qi, et al. Effects of environmental factors and maize growth on surface temperature to monitor soil water content[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2022, 38(20): 77-85. (in Chinese with English abstract)
Precision data screening and partition of field surface temperature based on the crop growth status
Xiao Chun’an, Cai Jiabing※, Chang Hongfang, Zhang Jingxiao, Xu Di
(1.,,, 100038,; 2., 100048,)
Field surface temperature is one of the most important parameters for the water/heat exchange between soil, crop and atmosphere, particularly for the remote sensing inversion model of irrigation areas. Among them, the crop canopy temperature and soil surface temperatureare often mixed in the field surface temperature data at the early growth stage, due to the crop growth and development in the row and plant spacing. Continuous observation can normally be implemented using the thermal infrared sensor of the automatic monitoring system. The mean value of monitored temperature data is usually used to replace the temperature at the actual position at present. The mixed temperature data can pose a great challenge to the calculation accuracy of the fine field irrigation model during data processing. In this study, an improved screening was combined with the Logistic crop growth model to accurately partition the massive monitoring data of field surface temperature, considering the Leaf Area Index (LAI), crop canopy height, and the key points of crop growth status. The measured temperature data of maize and sunflower was collected in the Yongji experimental station in Inner Mongolia of China in 2021. The scanning temperature data was obtained using the field monitoring system (CTMS-On line). The screening algorithm was then designed and verified. The field observation data of maize and sunflower was collected in the Jiefangzha irrigation field in 2015, while the maize data was in the Changchun experimental station of Jilin Province from 2018 to 2019. Results showed that: 1) An efficient determination was achieved in the data screening for the surface temperature of sparse vegetation in the fields. A logistic model was used to simulate the key points in the screening algorithm, considering the crop growth indicators of LAI and crop canopy height. 2) Taking the relative error as an example, the optimization ranges of canopy temperature and soil surface temperature were about 10 percentage points, and more than 5 percentage points, compared with the temperature measured by the hand-held thermometer. A higher accuracy of data screening was achieved in the canopy temperature and soil surface temperature acquisition. 3) The correction factor after the screening was then determined, according to the crop planting density and LAI. Among them, the correction factors of crop canopy temperature (0.9) and soil surface temperature (1.1) were selected for the maize. The correction factors for the sunflower were specified as the correction factors of crop canopy temperature of 0.7 and the correction factors of soil surface temperature of 1.2, due to the baseline of maximal LAI of 4. Therefore, one recommendation was proposed to apply the screening in different field situations. Specifically, each increasing value can increase the correction factors of crop canopy temperature by 0.35 and reduce the correction factors of soil surface temperature by 0.18 per increase of sunflower maximal LAI. Therefore, important technical support can be obtained for precision irrigation management for the better performance of field monitoring data. The finding can also provide a strong reference to deal with the field temperature data of sparse vegetation crops. A great contribution can then be made to the precision screening of remote sensing data from unmanned aerial vehicles and satellites.
temperature; sensors; soils; canopy; data screening; Logistic model
10.11975/j.issn.1002-6819.2022.22.010
S274;S513
A
1002-6819(2022)-22-0089-13
肖春安,蔡甲冰,常宏芳,等. 考慮作物生長(zhǎng)狀態(tài)的農(nóng)田表面溫度數(shù)據(jù)精量甄別與區(qū)分[J]. 農(nóng)業(yè)工程學(xué)報(bào),2022,38(22):89-101.doi:10.11975/j.issn.1002-6819.2022.22.010 http://www.tcsae.org
Xiao Chun’an, Cai Jiabing, Chang Hongfang, et al. Precision data screening and partition of field surface temperature based on the crop growth status[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2022, 38(22): 89-101. (in Chinese with English abstract) doi:10.11975/j.issn.1002-6819.2022.22.010 http://www.tcsae.org
2022-08-10
2022-10-10
國(guó)家自然科學(xué)基金項(xiàng)目(51979286;52130906);院地合作研究項(xiàng)目(HBAT02242202010-CG)資助
肖春安,研究方向?yàn)楣鄥^(qū)灌溉用水管理理論與技術(shù)。Email:xca1998@163.com
蔡甲冰,博士,教授級(jí)高級(jí)工程師,研究方向?yàn)楣鄥^(qū)灌溉用水管理理論與技術(shù)。Email:caijb@iwhr.com