remote sensing

8 downloads 0 Views 14MB Size Report
Dec 1, 2017 - Mechanically Sown Wheat with UAV RGB Imagery ... missing seedling regions were 12.39 cm for drill sowing and 0.20 cm2 for broadcast ...
remote sensing Article

Evaluation of Seed Emergence Uniformity of Mechanically Sown Wheat with UAV RGB Imagery Tao Liu 1 , Rui Li 1 , Xiuliang Jin 2 and Wenshan Guo 1, * 1

2

*

ID

, Jinfeng Ding 1 , Xinkai Zhu 1 , Chengming Sun 1, *

Jiangsu Key Laboratory of Crop Genetics and Physiology/Co-Innovation Center for Modern Production Technology of Grain Crops, Yangzhou University, Yangzhou 225009, China; [email protected] (T.L.); [email protected] (R.L.); [email protected] (J.D.); [email protected] (X.Z.) INRA-EMMAH, UMT-CAPTE, 84914 Avignon, France; [email protected] Correspondence: [email protected] (C.S.); [email protected] (W.G.)

Received: 29 September 2017; Accepted: 28 November 2017; Published: 1 December 2017

Abstract: The uniformity of wheat seed emergence is an important characteristic used to evaluate cultivars, cultivation mode and field management. Currently, researchers typically investigated the uniformity of seed emergence by manual measurement, a time-consuming and laborious process. This study employed field RGB images from unmanned aerial vehicles (UAVs) to obtain information related to the uniformity of wheat seed emergence and missing seedlings. The calculation of the length of areas with missing seedlings in both drill and broadcast sowing can be achieved by using an area localization algorithm, which facilitated the comprehensive evaluation of uniformity of seed emergence. Through a comparison between UAV images and the results of manual surveys used to gather data on the uniformity of seed emergence, the root-mean-square error (RMSE) was 0.44 for broadcast sowing and 0.64 for drill sowing. The RMSEs of the numbers of missing seedling regions for broadcast and drill sowing were 1.39 and 3.99, respectively. The RMSEs of the lengths of the missing seedling regions were 12.39 cm for drill sowing and 0.20 cm2 for broadcast sowing. The UAV image-based method provided a new and greatly improved method for efficiently measuring the uniformity of wheat seed emergence. The proposed method could provide a guideline for the intelligent evaluation of the uniformity of wheat seed emergence. Keywords: wheat; UAV image; uniformity; seedling-missing region; length of region of missing seedlings; area of region of missing seedlings; evaluation

1. Introduction Wheat seedlings grown under non-uniform conditions will result in yield loss. Excessive planting density will reduce the availability of nutrients and intensify intraspecific competition between plants, which will further affect growth and development as well as reduce productivity [1]. However, while a sparse planting density can ensure full development of individual plants, poor soil fertility and a lack of luminous energy may make it difficult to achieve ideal output, and result in different levels of reduced production. Many factors affect the uniformity of wheat seed emergence. Some problems with seed emergence are also research hotspots, such as the effects of straw mulching [2,3], as well as the effects of temperature, moisture and sowing depth [4,5]. However, most current research on the uniformity of wheat seed emergence involves manually selecting several regions, investigating the amount of plants in those regions, and calculating the coefficients of variation of the plants in those regions to reflect the uniformity of seed emergence. Such method for measuring the uniformity of seed emergence are time-consuming and laborious as well as inadequate because it cannot directly reflect the seed emergence conditions in a field scale.

Remote Sens. 2017, 9, 1241; doi:10.3390/rs9121241

www.mdpi.com/journal/remotesensing

Remote Sens. 2017, 9, 1241

2 of 15

Currently, UAV remote sensing is playing an important role in crop monitoring and features high efficiency, high resolution, low cost, low risk, flexibility and other characteristics [6–8]. Remote Sens. 2017, 9, 1241 2 of 15 Previous research studies have found that the application of UAV remote sensing is mainly used for crop classification, monitoring, plants density, estimation ofmonitoring agronomicand parameters Currently,disaster UAV remote sensing wheat is playing an important role in crop features and growth analysis in agricultural applications [9,10]. crop classification research, UAVs mainly acquire high efficiency, high resolution, low cost, low risk,In flexibility and other characteristics [6–8]. Previous research studies have found that thetoapplication of UAV remote sensing is mainly used for crop [11]. data on different types of planting areas facilitate crop management and related decision-making classification, disaster monitoring, wheat plants estimation of agronomic and of Disaster monitoring has been an important topicdensity, in recent years, including bothparameters the monitoring growth analysis in agricultural applications [9,10] In crop classification research, UAVs mainly field pests and the abnormal growth of crops caused by external factors, such as the monitoring of acquire data on different types of planting areas to facilitate crop management and related decisionlodging [12,13]. The plants density of wheat was estimated using image processing algorithm and making [11]. Disaster monitoring has been an important topic in recent years, including both the RGB imagery [14]. monitoring of field pests and the abnormal growth of crops caused by external factors, such as the The estimation of the agronomic parameters includes growth traits processing including leaf monitoring of lodging [12,13]. The plants density of mainly wheat was estimated using image area algorithm index, biomass, nitrogen content and plant height through the establishment of models. and RGB imagery [14]. Then, researchers looked for UAV image parameters that could reflect the quality of crop population The estimation of the agronomic parameters mainly includes growth traits including leaf area to index, content and plant heightmany through the establishment models. Then, evaluate thebiomass, status ofnitrogen crop growth [15,16]. Although applications of UAVsofare available related researchers looked for UAV image parameters could the quality of crop population to crop growth and management, few reports werethat about thereflect application of UAV remote sensingtoto the evaluate of crop of growth [15,16]. many applications of UAVs are available related evaluation of the thestatus uniformity wheat seed Although emergence. to crop growth and management, few reports were about the application UAVemergence remote sensing to the This study used UAV images to evaluate the uniformity of wheatofseed under the evaluation of the uniformity of wheat seed emergence. conditions of mechanical broadcast and mechanical drill sowing, to explore the evaluation methods This study used UAV images to evaluate the uniformity of wheat seed emergence under the and implementation theories of the study of the uniformity of seed emergence under different sowing conditions of mechanical broadcast and mechanical drill sowing, to explore the evaluation methods patterns. The goal was theories to provide support and of a seed theoretical basis for different the comprehensive and implementation of thetechnical study of the uniformity emergence under sowing evaluation of the wheattechnical seed emergence. patterns. The uniformity goal was to of provide support and a theoretical basis for the comprehensive evaluation of the uniformity of wheat seed emergence.

2. Materials and Methods

2. Materials and Methods

2.1. Field Trials

2.1. Field Trials

This study employed an experimental wheat variety, Yangfumai 4, with a seed emergence rate This studyemergence employed an experimental wheat Yangfumai 4,drill withand a seed emergencebroadcast rate of 80% and a field rate of 60% and usedvariety, both mechanical mechanical of 80% and a field emergence rate of 60% and used both mechanical drill and mechanical broadcast sowing patterns. Drill sowing machine put seeds in different rows, and the space between two lines sowing patterns. Drill sowingbetween machine seeds put seeds different and between two lines was 25 cm. The normal distance wasinabout 1 cmrows, to 2.5 cmthe inspace one line. Broadcast sowing was 25 cm. The normal distance between seeds was about 1 cm to 2.5 cm in one line. Broadcast sowing machine spread seeds on the field symmetrically, and the normal distance among seeds was about 6 cm machine spread seeds on the field symmetrically, and the normal distance among seeds was about 6 to 9 cm. Three planting densities were selected: 180 × 104 ha4 −1 ,−1240 × 1044 ha−1−1 , and 300 ×4 104−1 ha−1 cm to 9 cm. Three planting densities were selected: 180 × 10 ha , 240 × 10 ha , and 300 × 10 ha (Figure 1). The sizesize of each experiment wideand and3333mmlong. long. There were 6 plots (Figure 1). The of each experimentplot plot is is 2.7 2.7 m m wide There were 6 plots in thein the experiment: oneone replicate forfor each anddensity. density.Trials Trials were made in both experiment: replicate eachsowing sowingmethod method and were made in both 20142014 and and 2015.2015. The basic information of of thethe experiments inTable Table1.1. The basic information experiments is is shown shown in

Figure Theschematic schematic map map of Figure 1. 1.The ofthe theexperiment. experiment.

Remote Sens. 2017, 9, 1241

3 of 15

Table 1. Basic information of the experiments.

Remote Sens. 2017, 9, 1241

Year

Year 2014 2014 2015

2015

2.2. Image Acquisition

Sowing Date Soil of Characteristics (mg kg Table 1. Basic information the experiments.

Sowing Date 29 October 29 October 8 November

8 November

3 of 15

−1 )

hydrolysable N: 112.23 Soil Characteristics (mg kg−1) Available P: 45.61 hydrolysable AvailableN: K:112.23 137.16 Available P: 45.61 hydrolysable N: 99.86 Available K: 137.16 Available P: 39.12 hydrolysable N: 99.86 Available K: 106.55 Available P: 39.12 Available K: 106.55

2.2.DJI Image Acquisition A 1 Inspire RAW UAV was selected to obtain images. The camera calibration function of A DJI 1 Inspire RAW UAV was was selected to obtain images. The cameraItcalibration function of DJI DJI Go APP (http://www.dji.com/) used for camera calibration. was equipped with a RGB Go APP (http://www.dji.com/) was used for camera calibration. It was equipped with a RGB camera camera with 16 megapixels, which is capable of acquiring images vertically with a flight height of withspatial 16 megapixels, which is capable of acquiring images vertically a flight height of 7 mmicroplot (the 7 m (the resolution is about 0.6 mm). The coordinates of thewith subsample over each spatial resolution is about 0.6 mm). The coordinates of the subsample over each microplot were were manually defined using the orthomosaic image generated by the Agisoft Photoscan software manually defined using the orthomosaic image generated by the Agisoft Photoscan software (Version (Version 1.2.2, Agisoft LLC., St. Petersburg, Russia). Matlab (V2016a, MathWorks, Natick, MA, USA) 1.2.2, Agisoft LLC., St. Petersburg, Russia). Matlab (V2016a, MathWorks, Natick, MA, USA) was used was used to process the images. The flowchart of UAV image preprocessing is shown in Figure 2. to process the images. The flowchart of UAV image preprocessing is shown in Figure 2. In this study, In this study, wree on acquired on 172015 February and 26 February 2016. UVA data were UVA dataUVA wreedata acquired 17 February and 262015 February 2016. All UVA dataAll were acquired acquired during 3:00 p.m.–5:00 during 3:00 p.m.–5:00 p.m. p.m.

Figure 2. The flowchart of UAV (Unmanned Aerial Vehicle) image preprocessing.

Figure 2. The flowchart of UAV (Unmanned Aerial Vehicle) image preprocessing.

2.3. Image Processing

2.3. Image Processing

2.3.1. Image Segmentation

2.3.1. Image Segmentation

Based on the row spacing of drill sowing, the entire image in each date was divided into 30 cm

Based the row spacing ofand drillthen sowing, the entire in each datesub-region was divided × 30 cmon sub-regions (Figure 3), the percentage of image wheat cover in each was into Each replicate was divided to 990 percentage of was 30 cmcalculated. × 30 cm sub-regions (Figure 3), andinthen thesub-regions. percentageCoverage of wheat refers covertointhe each sub-region wheat seedling pixels in each sub-region. The image segmentation methods for drill and broadcast calculated. Each replicate was divided in to 990 sub-regions. Coverage refers to the percentage of werepixels the same. The vegetation index (Equation (1)) and Otsu’s algorithm usedand to extract wheatsowing seedling in each sub-region. The image segmentation methods were for drill broadcast wheat seedlings in the image [17]. The reference threshold value obtained by Otsu’s algorithm was sowing were the same. The vegetation index (Equation (1)) and Otsu’s algorithm were used to extract wheat seedlings in the image [17]. The reference threshold value obtained by Otsu’s algorithm was

Remote Sens. 2017, 9, 1241

4 of 15

Remote Sens. 2017, 9, 1241

4 of 15

about 40 in both dates. The rows of drill sowing should be detected before getting the missing seedling about 40 in both dates. The rows of drill sowing should be detected before getting the missing region. The missing seedling region of broadcast sowing could be obtained without detecting rows. seedling region. The missing seedling region of broadcast sowing could be obtained without Thus, different analytical methods were designed in this study. detecting rows. Thus, different analytical methods were designed in this study. Exg Exg==22×× G G− − RR −−BB

(1) (1)

The of the the three threecolor colorcomponents components(red, (red,green, green, and blue) RGB color images The values of and blue) in in RGB color images (24 (24 bits)bits) are, are, respectively, represented G, B. and B. respectively, represented as R,as G,R,and

Figure Figure 3. 3. Diagrammatic Diagrammatic sketch sketch of of sub-regions sub-regions in in an an image. image.

2.3.2. 2.3.2. Calculation Calculation of of Seed Seed Emergence Emergence Uniformity Uniformity The between sub-regions were calculated by Equations (2)–(4) and The coefficient coefficientofofvariation variation(CVs) (CVs) between sub-regions were calculated by Equations (2)–(4) the coverage of each sub-region. “x”, “i“, and “n” in Equations (2) and (3) are the coverage of each and the coverage of each sub-region. “x”, “i”, and “n” in Equations (2) and (3) are the coverage sub-region, the serial of each sub-region, and and the the total number of each sub-region, the number serial number of each sub-region, total numberofofeach each sub-region, sub-region, respectively. The coefficient of variation method is used to represent a uniformity index respectively. The coefficient of variation method is used to represent a uniformity index by by dividing dividing the standarddeviation deviation(δ(δ)x)byby the average value (EAx).smaller A smaller coefficient of variation indicates the standard the average value (E ). coefficient of variation indicates higher x x higher uniformity [18]. Equations (2)–(4) provide the calculations [19]: uniformity [18]. Equations (2)–(4) provide the calculations [19]: 1 =1 n Ex = n ∑ i =1 x i n r 11 n (( x − δx == − Ex)) n ∑ i =1 i 1 Ex Uni f ormity = 1 = =cv = δx cv 2.4. Localization of Seedlingless Ridges in Drill Sowing

(2) (2) (3) (3) (4) (4)

2.4. Localization of of Seedlingless Ridges in Drill Sowing Localization seedlingless ridges refers to seeking the condition of missing seedlings with drill sowing of wheat in the row direction, identifying the regions with missing seedlings, and calculating Localization of seedlingless ridges refers to seeking the condition of missing seedlings with drill the lengths of these regions. The localization steps were as follows. (1) Traverse the image based on sowing of wheat in the row direction, identifying the regions with missing seedlings, and calculating a 3 × 3 pixels (about 2 × 2 cm) template, extract the coverage of the template by using Equation (1) the lengths of these regions. The localization steps were as follows. (1) Traverse the image based on and Otsu’s algorithm, and construct a three-dimensional image of the area covered by the image. a 3 × 3 pixels (about 2 × 2 cm) template, extract the coverage of the template by using Equation (1) and Otsu’s algorithm, and construct a three-dimensional image of the area covered by the image. The three-dimensional image was constructed by coverage values (z-axis) and the coordinate (x-axis, y-

Remote Sens. 2017, 9, 1241

5 of 15

The three-dimensional image was constructed by coverage values (z-axis) and the coordinate (x-axis, y-axis) of the original image. In this type of image, the “hills” in a three-dimensional image represent the regions with wheat seedlings and the “cliff” areas in the row direction represent regions with missing seedlings. The regions of missing seedlings can be more intuitively seen through three-dimensional images. (2) Localize the wheat rows (the ridge in the three-dimensional image) using Equation (5). Each three-dimensional image is defined as I(m, k) , where m is the height of the image and k is the width of the image. (xi , yj ) is the coordinate of each pixel in the image, where Yj is the total value of each row’s coverage in the image. Rg is the center line of the wheat row which is obtained by Equation (6). (3) With row as the unit, calculate the change in horizontal coverage. (4) The portion where the mean value of the coverage change curve is equal to 0 is the region of missing seedlings. The length of the region of missing seedlings was calculated by adding all the length of 0 value points on the coverage change curve  m Yj = ∑i=0 xi , y j ( j = 1, 2, 3, · · · k) (5) Rg = Yj (Yj−1 < Yj ∩ Yj+1 < Yj )

(6)

2.5. Localization of Missing Seedling in Broadcast Sowing The extraction steps for locating regions of missing wheat seedlings with broadcast sowing included: (1) confirming the coordinates of the original image (this step was used to confirm the length and width); (2) traversing the image based on a 2 cm × 2 cm template and extracting the coverage of wheat in the template by using Equation (1) and Otsu’s algorithm; (3) extracting the regions whose value is 0 in the coverage distribution diagram and also with an area larger than 0.02 m2 ; and (4) counting the number and measuring the area of the regions of missing seedlings as well as collecting any relevant information from between the regions of missing seedlings. 2.6. Uniformity Evaluation The standard of evaluating uniformity was formulated by exploring the change of yield under different uniformity conditions. The field was divided into 2.5 m × 2.5 m regions. The grain yield and uniformity of each region was measured correspondingly. The grains were dried and weighed after manual harvesting, and then the grain yield of each region was obtained. The manual measurement steps of seed emergence uniformity are: (1) selecting 50 points randomly in each treatment; (2) marking each point by a 30 cm × 30 cm square and counting the seedling number manually; and (3) calculating the uniformity by Equations (2)–(4). The reference number of missing seedling region was counted manually. The length of missing seedling region for drill sowing field were measured by “Ruler” tool in the Agisoft Photoscan software. The area of missing seeding for broadcast sowing field were measured by “Polygon” tool in the Agisoft. The miss seedling rates of broadcast and drill sowing are calculated by Equation (7). MSR =

MS TS

(7)

MS is the length of missing seedlings for drill sowing, and is the area of missing seedlings for broadcast sowing. TS is the total length of wheat seedlings for drill sowing, and is the total area of wheat seedlings for broadcast sowing. 3. Results 3.1. Localization of Seedlingless Ridges in Drill Sowing Figure 4 shows the three-dimensional effect of drill sowing on wheat coverage. In the threedimensional diagram, wheat was clearly divided into different “hills.” The ridges can be easily positioned based on Equations (5) and (6).

Remote Sens. 2017, 9, 1241 Remote Sens. 2017, 9, 1241

6 of 15

Remote Sens. 2017, 9, 1241

6 of 15

6 of 15

Figure 4. The results of coverage calculation: (A) original Image of drill sowing wheat; and (B) threeFigure 4. The results of coverage calculation: (A) original Image of drill sowing wheat; and (B) threedimensional diagram of coverage. Figure 4. The resultsofofcoverage. coverage calculation: (A) original Image of drill sowing wheat; and (B) threedimensional diagram dimensional diagram of coverage.

Then, through the extraction value of the hill line, the method can clearly delineate the position throughseedlings the extraction value of the ridges hill line, the method delineate position of Then, the missing and seedlingless regions. Figurecan 5 clearly shows rangesthe values for of Then, through the extraction value of the hill line, the method can clearlythe delineate the position theextraction missing seedlings and seedlingless ridges regions. Figure 5 shows the ranges values for extraction of the first and fifth hill results in Figure 4. The result shows that the locations whose of the missing seedlings and seedlingless ridges regions. Figure 5 shows the ranges values for coverage values seedlingless regions. of the first and hill0 are results in Figure 4.ridge The shows theshows locations coverage values extraction offifth theare first andallfifth hill results in result Figure 4. Thethat result thatwhose the locations whose arecoverage 0 are all values seedlingless ridge regions. are 0 are all seedlingless ridge regions.

Figure 5. Changes in the coverage value of the first and fifth lines in Figure 4: (A) coverage value of the first line; and in (B)the coverage value of the fifth line. Figure 5. Changes lines in in Figure Figure4:4:(A) (A)coverage coverage value Figure 5. Changes in thecoverage coveragevalue valueof ofthe the first first and and fifth fifth lines value of of thethe first line; and (B) coverage value of the fifth line. first line; and (B) coverage value of the fifth line.

3.2. Localization of Missing Seedlings in Broadcast Sowing

Localization Missing Seedlings Broadcast Sowing 3.2.3.2. Localization of of Missing Seedlings Broadcast Figure 6 illustrates the processinin and resultsSowing of missing seedling localization in broadcast sowing. Through the extraction of the coverage and segmentation regions, the regions in of missing seedlings Figure 6 illustratesthe theprocess processand andresults results of of missing missingofseedling sowing. Figure 6 illustrates seedlinglocalization localization inbroadcast broadcast sowing. were accurately positioned. At the same time, the numberofofregions, regions of regions missing seedlings (RN), the Through extraction thecoverage coverage and segmentation seedlings Through thethe extraction ofofthe and segmentation of regions,the the regionsofofmissing missing seedlings area of each of these regions, the maximum and minimum area of these regions (MX and MN), and were accurately positioned. same time, number of regions of missing seedlings (RN), were accurately positioned. AtAt thethe same time, thethe number of regions of missing seedlings (RN), thethe area the mean value (MA) regions, and variance (ST) of theand areaminimum of these regions can beregions calculated according toand the area of each of these the maximum area of these (MX and MN), of characteristics each of these regions, the maximum and minimum area of these regions (MX and MN), and the of the regions. the mean and variance of the area theseregions regionscan can be be calculated thethe mean value value (MA)(MA) and variance (ST)(ST) of the area ofof these calculatedaccording accordingtoto characteristics of the regions. characteristics of the regions.

Remote Sens. 2017, 9, 1241

7 of 15

Remote Sens. 2017, 9, 1241

7 of 15

Remote Sens. 2017, 9, 1241

7 of 15

Figure 6. Extraction of regions withmissing missing seedlings seedlings for sowing of wheat: (A) Original Figure 6. Extraction of regions with forbroadcast broadcast sowing of wheat: (A) Original image of broadcast; (B) Color map of coverage distribution; (C) Missing seedling region; (D) image of broadcast; (B) Color map of coverage distribution; (C) Missing seedling region; (D) Information Figure 6. Extraction of of regions with missing seedlings for broadcast sowing of wheat: (A) Original Information extraction missing seedling region. extraction of missing seedling region. image of broadcast; (B) Color map of coverage distribution; (C) Missing seedling region; (D)

Information of missing seedling region. 3.2. Test of Wheatextraction Seed Emergence Uniformity

3.3. Test of Wheat Seed Emergence Uniformity

The summary of the measured uniformity by UAV is shown in Table 2. The test results of UAV 3.2. Test of Wheat Seed Emergence Uniformity image-based uniformity delineationuniformity of wheat seed is shown in Figure 7. From 1:1 line, The summary of the measured by emergence UAV is shown in Table 2. The test the results of UAV The the measuredofuniformity by UAV isofshown Table 2. Figure The test7. results UAV it can besummary seen thatofthe uniformity measurement value different sowing methods only of slightly image-based uniformity delineation wheat seed emergence is in shown in From the 1:1 line, 2 values image-based uniformity delineation of wheat seed emergence isThe shown in Figure 7. From the 1:1 line, varied when compared with the manual measurement value. R of drill and broadcast it can be seen that the uniformity measurement value of different sowing methods only slightly varied it can wheat be seen that0.83** the uniformity measurement value of different sowing methods slightly were and 0.89**, respectively. The RMSE 0.44and andbroadcast 0.64, only respectively. whensown compared with the manual measurement value. The R2values valueswere of drill sown wheat 2 values of drill and broadcast varied when compared with the manual measurement value. The R The maximum deviation of broadcast and drill sown wheat were 1.22 and 2.24, respectively. weresown 0.83**wheat and 0.89**, respectively. The RMSE values were 0.44 and 0.64, respectively. The maximum were 0.83** and 0.89**, respectively. The RMSE values were 0.44 and 0.64, respectively. deviation of broadcast and drill sown wheat were 1.22 and were 2.24, 1.22 respectively. The maximum deviation of broadcast and drill sown wheat and 2.24, respectively.

Figure 7. Test of unmanned aerial vehicle image-based wheat seed emergence uniformity: (A) broadcast sowing of wheat; and (B) drill sowing of wheat. Figure 7. of Test of unmanned vehicle image-based seed emergence uniformity: (A) Figure 7. Test unmanned aerial aerial vehicle image-based wheatwheat seed emergence uniformity: (A) broadcast broadcast sowing of wheat; and (B) drill sowing of wheat.

sowing of wheat; and (B) drill sowing of wheat.

Remote Sens. 2017, 9, 1241

8 of 15

Remote Sens. 2017, 9, 1241

8 of 15

Table 2. Summary of the measured uniformity by UAV. Table 2. Summary of the measured uniformity by UAV. Sowing Patterns Sowing Patterns Broadcast Broadcast drill drill

Nb. Subsamples Nb. Subsamples 120 120 120 120 1

1

Min Min 0.67 0.67 0.65

0.65

1 Max Mean Mean Range Range SD Max SD 1 5.2 2.64 4.53 0.99 5.2 2.64 4.53 0.99 6.22 3.16 5.57 1.66 6.22 3.16 5.57 1.66

SD, standard deviation.

SD, standard deviation.

3.4. 3.3. Test Test of of Drill Drill Sowing Sowing Wheat Wheat Seedling-Missing Seedling-Missing Region Region The summary summary of of the measured measured information information of of missing missing seedlings seedlings is is shown shown in in Table Table 3. 3. The The 1:1 line between the UAV UAV and and manual manual measurement measurement results results of of the the number number and and length length of drill sowing region of missing missing wheat wheatseedlings seedlingsisisshown shown Figure 8. The R2 value the number this region was inin Figure 8. The R2 value of theofnumber of thisofregion was 0.84**, 2 2 0.84**, and was RMSE was the Rof value of the of this region was and RMSE 3.99; the3.99; R value the length oflength this region was 0.61**, and0.61**, RMSEand wasRMSE 12.39. was The 12.39. Thenumbers average numbers of regions of missing seedlings obtained from UAVimages imagesand and manual average of regions of missing seedlings obtained from UAV measurement were 22.3 and 23.2, respectively. The average length of these regions obtained from UAV difference. and manual measurement measurement was was 38.7 38.7cm cmand and43.9 43.9cm, cm,respectively, respectively,which whichshows showsonly onlya small a small difference.

Figure 8. Test results of the: number (A); and length (B) of regions with missing seedlings for drill Figure 8. Test results of the: number (A); and length (B) of regions with missing seedlings for drill sowing of wheat regions with missing seedlings. Note: NMSR and LMSR refer to the number and sowing of wheat regions with missing seedlings. Note: NMSR and LMSR refer to the number and length (cm) of regions with missing seedlings, respectively. length (cm) of regions with missing seedlings, respectively. Table 3. Summary of the measured the number and length of missing seedlings for drill sowing by UAV. Table 3. Summary of the measured the number and length of missing seedlings for drill sowing by UAV. Info. of MSR Nb. Subsamples Min Max Mean Range SD

Number

Info. of MSR Length (cm)

30

Nb. Subsamples 30

6

Min 11

40 22.27 Max 38.67 Mean 101

Number 30 6 40 Length (cm) 30 11 101 3.4. Test of Broadcast Sowing Wheat Seedling-Missing Region

22.27 38.67

34

9.67

Range SD 90 17.5 34 9.67 90 17.5

The summary of the measured information of missing seedlings is shown in Table 4. The 1:1 line 3.5. Test of Broadcast Sowing Wheat Seedling-Missing Region between the UAV measurement results and manual measurement results of the number and area of The summary of the of measured of missing seedlings is shown Table of 4. the Thenumber 1:1 line broadcast sowing region missing information wheat seedlings is shown in Figure 9. The Rin2 value 2 between UAV measurement results andand manual of the number and area of of regionthe of missing seedlings was 0.93**, RMSEmeasurement was 1.39; theresults R value of the area of region 2 value of the number broadcast sowing region of missing shown in Figure 9. The missing seedlings was 0.78**, and wheat RMSE seedlings was 0.20.isThe average number of R these regions obtained of region missing seedlings 0.93**, andwas RMSE was5.2, 1.39; the R2 value of average the areaarea of region of from UAVofimages and manual was measurement 4.5 and respectively. The of these 2, respectively, missing seedlingsfrom was 0.78**, andmanual RMSE was 0.20. The average these obtainedwhich from regions obtained UAV and measurement was 0.57number cm2 andof0.51 cmregions UAV images and manual measurement was 4.5 and 5.2, respectively. The average area of these regions shows little difference. obtained from UAV and manual measurement was 0.57 cm2 and 0.51 cm2 , respectively, which shows little difference.

Remote Sens. 2017, 9, 1241

9 of 15

Remote Sens. 2017, 9, 1241

9 of 15

Figure 9. Test results of the: number (A); and area (B) of regions with missing seedlings for

Figure 9. Test results of the: number (A); and area (B) of regions with missing seedlings for broadcast sowing Note: NMSR and LMSR refer to the number and area of regions with missing broadcast sowing Note: NMSR and LMSR refer to the number and area of regions with missing seedlings, respectively. seedlings, respectively. Table 4. Summary of the measured the number and area of missing seedlings for broadcast sowing

Tableby4.UAV. Summary of the measured the number and area of missing seedlings for broadcast sowing by UAV. Info. of MSR Nb. Subsamples Min Max Mean Range SD Number 30 0 17 5.2 17 4.2 Info. of MSR Nb. Subsamples Min Max Mean Range SD Area (m2) 50 0.01 0.97 0.51 0.96 0.37 Number Area (m2 )

30 50

0 0.01

3.5. Uniformity Evaluation under Different Plant Density

17 0.97

5.2 0.51

17 0.96

4.2 0.37

The present study found for both drillDensity and broadcast sown wheat, changes in the uniformity 3.6. Uniformity Evaluation under that, Different Plant of seedling emergence always had certain effects on yield (Table 5). At three different planting The present study that, both drill and broadcast sownthe wheat, changes in the uniformity densities, when the found measure of for uniformity was larger than four, population yield remained of seedling emergence always had certain effects on yield (Table 5). At three different planting densities, relatively high and stable. When the measure of uniformity fell between two and four, the yield moreof widely, and, when was lessthe than two, yield was relatively low.relatively high whenfluctuated the measure uniformity wasuniformity larger than four, population yield remained

and stable. When the measure of uniformity fell between two and four, the yield fluctuated more Table 5. Relationship between uniformity and yield (kg ha−1). widely, and, when uniformity was less than two, yield was relatively low. Density (104 ha−1) −1 Table Uniformity 5. Relationship between uniformity 180 240 and yield 300(kg ha ). >4 (U1) 2–4 (U2) Uniformity 4, 2–4, and6900–7300 4 and (U1)U3 indicate 6800–7100 7400–7800 2–4 (U2) 6100–7100 7100–7600 6200–7200 The length of a drill sowing region with missing seedlings had little effect on the yield in a certain 4, 2–4, and 20, respectively. 20 (L3) 5800–6000 6000–6500

drill300 sowing) was 4, 2–4, and