4309
Geotagged Application for Durian Trees using Aerial Imagery and Vegetation Indices Algorithm
Mohidem, N. A.1, Jeya Kumaran, V.2, Ab Rahman, M. I.3, Che’Ya, N. N.4*, Rosle, R.5, Fazlil Ilahi, W. F.6, Hock, O. C.7, Mat Su, A. S.8, Ping, T. N.9, Izan, S.10
Department of Agriculture Technology, Faculty of Agriculture, Universiti Putra Malaysia, 43400, UPM Serdang, Selangor, Malaysia1,2,3,4,5,6,8
Department of Land Management, Faculty of Agriculture, Universiti Putra Malaysia, 43400 Selangor, Malaysia9
Department of Crop Science, Faculty of Agriculture, Universiti Putra Malaysia, 43400 Selangor, Malaysia10 Institute Plantation Studies, Universiti Putra Malaysia, 43400 Selangor, Malaysia4,8
Marvelane Sdn. Bhd., 19, Jalan Opera F U2/F, Taman TTDI Jaya, 40150 Shah Alam, Selangor7 Corressponding Author: 4*
Keywords: ABSTRACT
Aerial imagery, Durian, GIS, Unmanned aerial vehicles, Vegetation index.
Durian demand has increased considerably, and it has gained popularity in the market. Under Industrial Revolution 4.0, precision agriculture is expanding globally with a wide range of digital technologies that provide the farming industry with information to improve farm productivity. The objectives of this study are to geotag the durian trees and to compare several Vegetation Indices (VIs) algorithms (Visible- Band Difference Vegetation Index (VDVI), Visible Atmospherically Resistant Index (VARI), Normalized Green-Red Difference Index (NGRDI), Red-Green Ratio Index (RGRI), Modified Green-Red Vegetation Index (MGRVI), Excess Green Index (ExG), Color Index of Vegetation (CIVE), and Vegetativen (VEG)). One hundred sixty durian trees at the Durian Valley in Kluang (Johor), were tagged, which consist of four sample trees for each treatment. Every two weeks of ground data such as the height of trees, canopy width, girth’s diameter, node distance, pH value, moisture content, electrical conductivity (EC) reading, and leaf sizes were exported into the QGIS software and joined with the tagged durian trees. The aerial imagery data captured the durian plantation area using Red Green Blue (RGB) sensor with a 100 m flight attitude. pH, EC, and moisture content were interpolated using Inverse Distance Weighted (IDW) technique. The processed image by VIs and geotagged trees could help farmers to identify the problem areas in the farm and monitor durian plantation effectively.
This work is licensed under a Creative Commons Attribution Non-Commercial 4.0 International License.
1. INTRODUCTION
Fruit is one of the food sources that provide many nutrients to humans. Durian or durio zibethinus is one
4310 type of favoured and expensive fruit, which is generally known as the King of Fruit. It is a tropical fruit in which various tree species belong to the order Malvales (family Malvaceae). The common name ‘durian’
derived from the Malay word ‘duri’ that means thorn, and the species name ‘zibethinus’ comes from the Italian word ‘zibeto’ that means strong aroma [1]. It has a delicious taste with a creamy texture and a sweet- bitter taste. Durian is a popular fruit and essential economic crop in Southeast Asian countries such as Malaysia, Thailand, and Indonesia. [2] estimated that the quantity of domestic demand is around 5,000 to 10,000 tons per year, of which 90% is imported from Thailand and the rest from Malaysia. Profitable durian plantations from these two countries can produce 10 to 18 tons per hectares.
Durian is the largest fruit plantation area in Malaysia, consisting of 72,536 hectares. It is mainly located in the states of Johor, Pahang, and Sarawak, and Musang King, Tekka, Duri Hitam and Udang Merah are examples of durian types in Malaysia. The production quantity of durian has increased since 2017, up to 348,170 metric tons in 2019 [3]. The Malaysian industry has exported durian fruits to neighbouring countries such as Singapore, Indonesia, Brunei, and other Asian countries such as China [4]. During the virus corona (COVID-19) pandemic, the Ministry of Agriculture and Food Industry (MAFI) aimed to export Musang King to China in 2020, which was expected to bring RM50 million worth. The normal retail cost for the durian Musang King was between RM40 to RM90 per kg in Malaysia, whereas in China, the durian can be sold at a cost between RM200 to RM400 per kg [5].
In a traditional approach, farmers typically use paper-based records to keep all the field data of durian trees.
This practice is impractical because it is time-consuming, subsequently enhanced to poor management. The crop monitoring is also challenging to handle the data that is not up to date. The paper-based record can limit data sharing, which leads to missing or inaccurate data. Hence, the conventional method is only reliance on visual observation to monitor the plant health. Nowadays, it is irrelevant to keep all the field data using paper- based records. As a solution, geotagging techniques could be used to tag the durian trees.
This technique includes adding geospatial data such as altitude, latitude, and longitude to various media shapes such as photos or videos. Farmers can capture the Global Positioning System location when the image is taken through a built-in camera in smartphones [6]. Farmers can conduct tree counting and geolocation of the trees from aerial geotagged images. Therefore, geotagging may be used to uniquely identify, track, and count the number of durian trees in a series of aerial images [7].
With this in mind, agricultural technology has changed from a traditional to a modern system, in accordance with the Industrial Revolution 4.0 (IR 4.0). Remote sensing technologies in precision agriculture could help the farmers to keep the field data effectively in a cost-effective and timely manner. One of the technologies is UAV or drone [8]. UAV can assist the farmers to identify crop problems through the atmosphere that could not be visualized at the ground level spot control [9]. The system is built to monitor the crop in farming activities such as growth, livestock conditions, irrigation system, application of fertilizers, and disease detection. UAV sensors with very high spectral and temporal resolutions were adequate for plant disease detection in the early stages [10]. It also helps to detect the crop health that can be tuned by the infrared sensor, which allows farmers to improve the crop conditions and react with insecticides or fertilizer [11].
Durian plantation mostly covers a large area and needs proper crop management. Hence, farmers can use the GIS application to store and manage spatial data. A geographic information system (GIS) is a tool composed of software, hardware, data, and users that can capture, store, manage, and analyze digital data such as graphs and maps as well as alphanumeric data [12]. According to Tongkaw [13], GIS can create a spatial database by the store the data to monitor the durian plantation. The GIS techniques can be carried
4311 out using open-source software such as the quantum geographical information system (QGIS), which supports many rasters, vector and database formats. QGIS also contains the tools for collecting and processing the data and designing and export maps.
The successful application of UAV depends on changes in sensitivity to VIs and growth stages [14]. Aerial imagery can extract VIs that help the farmer monitor the crop stress conditions and variability [15].
Vegetation indices are mathematical estimations of canopy reflectance at specified visible and near-infrared wavelengths [16]. Some VIs employ a narrow band of reflectance values in the electromagnetic spectrum.
Therefore, it will give more accurate measurements for crop yield and reliable information for yield prediction [17]. The indices also allow users to determine stressed and diseased crops and predict nitrogen (N) status by multiple bands imagery. Farmers with this knowledge can maximise the use of limited resources such as fertiliser, water, or manure by focusing on different parts of the field and conducting on- the-spot treatments. Accordingly, aerial imagery and vegetation indices data from the image processing technique could be exported, and trees and characteristics of the ground data could be plotted using GIS techniques. Then, this spatial data could be processed and displayed in the form of a map. GIS and aerial imagery approaches are an excellent combination to monitor the durian plantation and help the farmer detect the problem area precisely [18]. This study aimed to geotag the durian trees and compare VIs between VDVI, VARI, NGRDI, RGRI, MGRVI, ExG, CIVE, and VEG algorithms.
2. MATERIALS AND METHOD
2.1 STUDY AREA
The study was conducted in Kluang, Johor. Fig 1 (a) shows the geographic location is between longitude 103.32° East and latitude 2.03° North, and Fig 1 (b) shows the sampling area at the Durian Valley (a collaboration project with Marvelane Sdn Bhd). The total area of the study plot is about 7 acres. The Durian Valley has planted a variety of durian Musang King.
Fig. 1. (a) shows the location of Kuang, Johor in peninsular Malaysia, and (b) shows the sampling area at Durian Valley. The map was extracted from Google Earth.
2.2 GROUND DATA
4312 Ground data were collected every two weeks for the verification and validation of the imagery. This ground data consists of agronomic data of the durian trees such as the height of trees, canopy width, girth’s diameter, node distance, pH, measure content, EC reading, and leaf sizes. The plot has 160 trees with two years old. The selected trees were chosen for the ground data in the weekly record. Ground data were exported to a spatial database using QGIS software. The collection of attribute data in the software is defined as the data with features and contains the information of tree features. The data in the attribute table was joined with the points of geotagged trees.
2.3 AERIAL IMAGERY DATA
The aerial image data acquisition of the durian plantation area was captured using an RGB sensor. The UAV's function is to identify and monitor the crop of the durian plantation. Fig 2 shows that the outlook of the multirotor DJI Phantom 4. It is a small size that can fly for a short distance below 2 km from the control station in 100 m altitudes air space [16].
Fig. 2. DJI Phantom 4
2.4 SAMPLING DESIGN
There 160 durian trees were geotagged using QGIS software. Durian trees were joined with the attributes table of the software, which consists of ground data. Four sample trees were labelled as T1, T2, T3, and T4 for each treatment in each block. The sample trees selected for each treatment as follows: (i) T1: Tagged with number 7, 8, 9 & 10, (ii) T2: Tagged with number 7, 8, 9 & 10, (iii) T3: Tagged with number 7, 8, 9 &
10, and (iv) T4: Tagged with number 7, 8, 9 & 10.
2.5 FLIGHT PLANNING
The UAV was flying on 3rd January 2020. The flight attitude of the UAV was 100 m, which produced a 2.47 cm resolution. A remote controller controlled the UAV. Waypoint was designed to produce a comprehensive overlap map. The flight mission was preprogrammed and planned in the flight mission planner software. The front lap and side lap distances were set based on the total coverage area. There was an extra track or space for the UAV to stabilize for the straight track and complete coverage of the study plot. The calibrated image will prevent temporal and spatial analysis errors from occurring. The flight mission was carried out from 10:30 to 11:30 a.m., in which it did not take a long time to fly the UAV because the weather during that time was good. The wind speed was lower in the morning, and the captured image was free of the UAV's shadow.
2.5 IMAGE PROCESSING
The image processing used Agisoft Metashape software to mask the image. The workflow of image processing is shown in Fig 3. The raw image from the RGB cameras was converted into tiff file format using Agisoft Photoscan Professional 1.4.3. The first to final steps for mosaicking are as follows: (i) input raw image, (ii) align photo, (iii) build dense point cloud, (iv) build mesh, (v) build orthomosaic, and (vi)
4313 export orthomosaic image.
Fig. 3. Workflow of image processing 3. RESULTS AND DISCUSSION
3.1 VEGETATION INDICES
The vegetation index VIs that are used in the image processing are presented in Table 1. The data acquisition is generated with equations following their VI. The aerial imagery data was then imported to QGIS software to produce the vegetation index maps.
TABLE 1: Vegetation indices derived from RGB images
Vegetation Indices Formula Referenc
es Visible-band Difference Vegetation
Index (VDVI)
(2 ∗ 𝐺 − 𝑅 − 𝐵) (2 ∗ 𝐺 + 𝑅 + 𝐵)
[19]
Visible Atmospherically Resistant Index
(VARI)
(𝐺 − 𝑅) (𝐺 + 𝑅 − 𝐵)
[20]
Normalized Green-Red Difference Index
(NGRDI)
𝐺 − 𝑅 𝐺 + 𝑅
[21]
Red-Green Ratio Index (RGRI) 𝑅
𝐺
[22]
Modified Green Red Vegetation Index (MGRVI)
(𝐺2 − 𝑅2) (𝐺2 − 𝑅2)
[23]
Excess Green Index (ExG) 2 ∗ 𝐺 − 𝑅 − 𝐵 [24]
Color Index of Vegetation (CIVE) 0.441 ∗ 𝑟 − 0.881 ∗ 𝐺 + 0.385𝐵 + 18.787
[25]
Vegetativen (VEG) 𝐺
(𝑅)𝑎∗ ∗ 𝐵(1−𝑎)
[26]
3.2 SPATIAL ANALYSIS
Contour, water flow, and digital elevation model (DEM) were represented in the QGIS. This study also used Inverse Distance Weighted (IDW) to interpolate the values of pH, EC, and moisture content. The measured values near the prediction area are applied to predict a value for any unsampled area, assuming that things closest to one another are more similar than things that are farther apart [27]. Finally, contour, water flow, DEM, and interpolated map were displayed using QGIS software.
4314 3.3 CORRELATION ANALYSIS
The correlation analysis was used to compare the canopy with the VIs algorithms such as VDVI, VARI, NGRDI, RGRI, MGRVI, ExG, CIVE, and VEG. The ground truth canopy was used to examine the relationship between several vegetation indices and ground truth canopy.
3.4 ACCURACY ASSESSMENT OF CANOPY DIAMETER
The accuracy of canopy diameter for durian trees were measured by comparing the results obtained between ground truth observation of 48 trees and the virtual surveyor. The canopy diameter was measured at the ground. Then the values were compared with the diameter of the canopy in the aerial imagery. The formula for the accuracy of vegetation analysis is shown in equation 1:
𝐴𝑐𝑐𝑢𝑟𝑎𝑐𝑦 𝑜𝑓 𝑣𝑒𝑔𝑒𝑡𝑎𝑡𝑖𝑜𝑛 𝑎𝑛𝑎𝑙𝑦𝑠𝑖𝑠 =𝑇𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡𝑟𝑒𝑠𝑠 𝑝𝑟𝑒𝑑𝑖𝑐𝑡𝑒𝑑 𝑐𝑜𝑟𝑟𝑒𝑐𝑡𝑙𝑦 𝑇𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡𝑟𝑒𝑠𝑠 𝑥100
3.5 JOINED THE GEO-TAGGED OF DURIAN TREES WITH GROUND DATA
Fig 4 shows the tagged durian tree with a unique code number. Geotagging is proven in visualizing geographical information such as longitude, latitude and altitude, and ground data of the durian tree. The geotagging technique is important because the mosaicked image created from inaccurately geotagged individual images could display an ambiguous time scene [28]. Also, image geotagging is essential for proper scaling and georeferencing of the related 3D photogrammetry mapping. [29]. In the future, this technique could be used by adding metadata of photos, video files, and audio files for durian and other crop types on the farm.
Fig. 4. Tagged of durian trees with a code number
Fig 5 shows the geotagged of durian trees data that was spatially joined with its ground data in the attribute table of QGIS software, and Fig 6 shows the screenshot of the attribute table in the software. The characteristics of ground data compiled in the spatial database enable farmers to effectively monitor the location of each tree.
4315 Fig. 5. Geotagged of durian trees that joined with its ground data
3.6 PROCESSED IMAGE
Orthomosaics image was created by integrating all of the images from the boundary area of the farm. The classification and segmentation of orthomosaic images for the extraction of pixels assigned and identified as durian trees. In Fig 5, the RGB orthomosaic image of durian trees at the sampling sites were marked with points in blue, pink, and yellow colours. The quality resolution of acquired image data is necessary and needs careful consideration in order to generate findings with a high degree of accuracy. Consequently, obtaining images with high spectral information of the entire field by orthomosaicing is a necessary step in initiating the detection process [30]. Farmers can use the orthomosaics to identify areas that need crop management, fertilizer application, insect control, and other treatments [31].
3.7 VEGETATION INDEX MAP
Based on the generated orthomosaics, the VIs including VDVI, VARI, NGRDI, RGRI, MGRVI, ExG, CIVE, and VEG were analysed. The created VIs map visualized chlorophyll content in the durian plantation area. VDVI images are visible light remote sensing images that can differentiate between vegetation and non- vegetation areas without employing the near-infrared (NIR) domain. The range of VDVI values are within −1 to 1. The red colour represents value of -1 which is less, or no vegetation and the green colour represents value of 1, which is weed and durian trees zones. The accuracy of VDVI-based vegetation extraction is higher than other visible light band-based VIs and G bands [32]. VARI is designed to emphasize vegetation in the visible portion of the spectrum. The red colour represents value of -9 which is less or no vegetation and the green colour represents the value of 19.50, which is the vegetation zone. VARI can minimize the sensitivity of the atmospheric effects and allows estimation of the vegetation fraction.
The NGRDI is an index that utilizes the normalized ratio of the difference between the green and red bands to exclude the effect of the different irradiance on the spectral properties of vegetation. The red colour represents value is -1 which has less or no vegetation and the green colour represents value is 1, which is the vegetation zone. The RGRI index is used to estimate and classify divergent patterns of pigment activity such as anthocyanin, which is closely related to light regimes, leaf structure, and functional type. The red colour represents a value of 0.00, which is no vegetation, and the green colour represents a value of 6.67,
4316 which is the least vegetation zone. MGRVI is the normalized difference between the squared green reflectance and the squared red reflectance. The red colour represents value is -1, which means less vegetation or soil and the green colour represents value is 1, which is the vegetation zone. ExG is the green portion of the electromagnetic spectrum to highlight the green vegetation cover. The red colour represents the value of -50, which is less or no vegetation, and the green colour represents the value of 205, which is the green vegetation cover zone.
CIVE is the index that separates the plant from the soil background. The red colour represents the value of - 77.11, which is less or no vegetation full of soil, and the green colour represents the value of 37.38, which is the vegetation zone. Moreover, the index outperforms the Near-infrared (NIR) technique in terms of plant segmentation since it places a greater emphasis on green areas. CIVE can increase the green information in images, allowing the green plant part to be separated from its background. Hence, CIVE layers were used as input to separate vegetation from the soil. VEG is the index that has been used to identify plant pixels. The red colour represents a value of 0.44, which is less or no vegetation, and the green colour represents the value of 32.00, which is more of a vegetation zone. VEG was suggested as a technique for separating the green component using an index that is constant across the range of natural daylight.
Fig 6 shows the Vis value visualization map, which consist of VDVI map (a), VARI map (b), NGRDI map (c), RGRI map (d), RGRI map (e), ExG map (f), CIVE map (g), and VEG map (h). One of the obvious benefits of using VIs is extracting the green colour channel successfully and then applying some specific thresholding techniques, such as (Otsu threshold), to achieve the desired findings. If we are using binary images, this will be a little easier because some binary segmentation algorithms will do the task successfully. To obtain the maximum accuracy from this method, crop images should be acquired precisely in the early season of the crop so that immediately colour-based segmentation may be applied to segment weed patches and actual crops by utilizing any of the colour indexes [33].
(a) VDVI map (b) VARI map
4317 (c) NGRDI value visualization map (d) RGRI value visualization map
(e) MGRVI value visualization map (f) ExG value visualization map
(g) CIVE value visualization map (h) VEG value visualization map
Fig. 6. Vegetation indices (Vis) value visualization map for each algorithm.
3.8 SPATIAL ANALYSIS
Slope can be measured at the study area based on the mechanical characteristics of the soil. Then, the slope can be determined by the stability calculation that identifies the failure curve, in which the slip risk is highest and the corresponding value of the safety factor [34]. The topography map is a bit hilly and not a flat area. This is because the original crop that was planted in this area was a rubber tree, then the owner converted it into the durian plantation as a new business. Fig 7 shows the water flow of the durian plantation. The yellow arrows show the direction of the water flow in the area. The water will go down to the lower area and based on the map. Thus, farmers can identify the suitable area to irrigate and put higher pressure on the irrigation system. A digital elevation model (DEM) is a 3D representation of a terrain's surface generated from elevation data. DEM is commonly used to refer to any digital representation of a topographic surface [35].
4318 Fig. 7. Water flow of durian plantation.
From the interpolation of pH values, the yellow color represents the value of 6.18 because of the effect on the increasing height of the durian trees. The black color presents the value of 5.61, which is acidic soil but also gives the effect of the durian height trees (Fig 8(a)). Additionally, from the EC interpolation, the blue color represents the value of 0.22, which is slightly saline that needs additional water to flush the excessive slats below the root zone during irrigation and the red color represent the value of 0.101 which non-saline area that less of salts of soil (Fig 8(b)). On the other hand, from the moisture content interpolation, the yellow color represents the value of 24.84, and the purple color represents the value of 10.57 (Fig 8(c)).
4319 Fig. 8. Interpolation of pH (a), EC (b), and moisture content (c).
3.9 ACCURACY OF CANOPY DIAMETER
The positions of the durian trees have been validated with aerial imagery and imported to virtual surveyor software. The accuracy of the canopy diameter obtained with the ground truth is 61%. Table 2 shows the comparison of the canopy diameter between the Virtual Surveyor software and ground truth observation.
This accuracy can be improved with the high-resolution imagery, and this method can help farmers in their plantations using UAV and GIS applications.
TABLE 2: Comparison of the canopy from the virtual surveyor with the ground measure ID Virtual
surveyor canopy measure (m)
Ground canopy measure
(m)
ID Virtual surveyor
canopy measure (m)
Ground canopy measure (m)
ID Virtual surveyor
canopy measure
(m)
Ground canopy measure
(m)
4B1T1 0.84 1.96 1B2T1 1.43 2.88 7B3T1 2.7 3.23
5B1T1 1.53 3.22 2B2T1 2.12 3.28 8B3T1 1.48 2.57
6B1T1 1.4 2.75 3B2T1 1.76 2.25 9B3T1 2 2.82
7B1T1 1.5 2.74 4B2T1 0.81 2.04 10B3T1 1.36 3.34
1B1T2 2.6 3.55 3B2T2 1.41 1.63 7B3T2 2.23 2.96
2B1T2 1.92 2.57 5B2T2 1.6 2.61 8B3T2 0.88 3.41
3B1T2 1.71 2.36 7B2T2 1.3 2.25 9B3T2 2.4 1.94
4B1T2 2.5 2.57 9B2T2 1.04 3.13 10B3T2 1.51 3.48
4B1T3 2.25 3.12 3B2T3 0.85 2.22 1B3T3 1.51 2.38
5B1T3 2.07 2.69 4B2T3 0.83 1.92 2B3T3 2.5 3.65
6B1T3 2.12 3.65 5B2T3 1.12 3.06 3B3T3 1.41 2.05
7B1T3 0.94 3.25 6B2T3 2.5 2.99 4B3T3 1.15 1.53
2B1T4 2.08 3.14 5B2T4 1 2.4 1B3T4 2.49 2.99
3B1T4 2.18 2.78 7B2T4 1.07 2.25 2B3T4 2.11 2.74
4B1T4 1.62 2.45 8B2T4 0.96 2.08 3B3T4 2.13 2.9
5B1T4 2.24 3.15 10B2T4 1.53 2.37 4B3T4 0.36 1.6
4. CONCLUSION
This study effectively demonstrated the potential of UAV as one of the precision agriculture tools to
4320 monitor the durian trees. This technology enhances the detection of the problem area quickly, rather than the conventional method, which is time-consuming and needs to monitor crops from the ground.
Geotagging is a technique that can help farmers monitor the durian plantation effectively by searching for information on each tree based on the geotag ID. Furthermore, the processed image that VIs help farmers analyse to make crop management decisions more efficiently. The benefits of UAVs were easily obtained from the market, considered cost-effective in the field application, and can produce a high spatial resolution image. Therefore, it is suitable for the farmers to use because they can monitor every part of the durian trees, especially in a large plantation area.
Research funding: The authors thank Universiti Putra Malaysia for providing financial support through Geran Putra UPM.RMC.800/2/2/4-Geran Putra (Vot No. 9693300): Development of geotagging system and crop monitoring using aerial imagery and vegetation indices algorithm. We would like to give Special thanks to MARLELANE Sdn Bhd and Kluang Valley Durian for providing the research area.
Author contributions: All authors have accepted responsibility for the entire content of this manuscript and approved its submission.
Competing interests: We declare no competing interests.
5. REFERENCES
[1] Thongkum, M., McAtee, P. M., Schaffer, R. J., Allan, A. C., & Ketsa, S. (2018). Characterization and differential expression of ethylene receptor genes during fruit development and dehiscence of durian (Durio zibethinus). Scientia Horticulturae, 240, 623-630.
[2] Ketsa, S., Wisutiamonkul, A., Palapol, Y., & Paull, R. E. (2020). The durian: Botany, horticulture, and utilization. Horticultural Reviews, 47, 125-211.
[3] Ministry of Agriculture and Food Industries (MAFI). (2020). MITI Report 2020. Malaysia
[4] Ketsa, S., Wisutiamonkul, A., Palapol, Y., & Paull, R. E. (2020). The durian: Botany, horticulture, and utilization. Horticultural Reviews, 47, 125-211.
[5] Ministry of Agriculture and Food Industries (MAFI). (2020). MITI Report 2020. Malaysia
[6] Atunggal, D., Ausi, N.H., Ma’ruf, B., dan Rokhmana, C. A. (2017, Oktober 11-13). Application of Low Cost RTK GPS Module for Precise Geotagging using Smartphone. 9th Multi GNSS Asia, Jakarta, 89- 92.
[7] Ammar, A., Koubaa, A., & Benjdira, B. (2021). Deep-Learning-based Automated Palm Tree Counting and Geolocation in Large Farms from Aerial Geotagged Images. Agronomy, 11(8), 1458.
[8] Che'Ya, Nik & Fadzilah, M & Roslin, Nor & Zainol Abdullah, Wan Nor Zanariah & Tarmidi, Zakri. (2019). Unmanned Aerial Vehicle Applications In Agriculture. IOP Conference Series: Materials Science and Engineering. 506. 012063. 10.1088/1757-899X/506/1/012063.
[9] Jiao, W., Wang, L., & McCabe, M. F. (2021). Multi-sensor remote sensing for drought characterization: current status, opportunities and a roadmap for the future. Remote Sensing of
4321 Environment, 256, 112313.
[10] Su, J., Liu, C., Hu, X., Xu, X., Guo, L., & Chen, W. H. (2019). Spatio-temporal monitoring of wheat yellow rust using UAV multispectral imagery. Computers and electronics in agriculture, 167, 105035.
[11] Bhoi, P. B., Wali, V. S., Swain, D. K., Sharma, K., Bhoi, A. K., Bacco, M., & Barsocchi, P. (2021).
Input Use Efficiency Management for Paddy Production Systems in India: A Machine Learning Approach.
Agriculture, 11(9), 837.
[12] Jnana, I., Putra, I., & Mandenni, N. M. I. M. (2021). Web GIS-Based Water Distribution Network Management using Google Maps. Water and Energy International, 64(3), 6-12.
[13] Tongkaw, S. (2021). Management Information Systems and Geographic Information System for Managing Durian Resources. Annals of Emerging Technologies in Computing (AETiC), 5(5).
[14] Olson, D., Chatterjee, A., Franzen, D. W., & Day, S. S. (2019). Relationship of drone‐based vegetation indices with corn and sugarbeet yields. Agronomy Journal, 111(5), 2545-2557.
[15] Radoglou-Grammatikis, P., Sarigiannidis, P., Lagkas, T., & Moscholios, I. (2020). A compilation of UAV applications for precision agriculture. Computer Networks, 172, 107148.
[16] Frels, K., Guttieri, M., Joyce, B., Leavitt, B., & Baenziger, P. S. (2018). Evaluating canopy spectral reflectance vegetation indices to estimate nitrogen use traits in hard winter wheat. Field Crops Research, 217, 82-92.
[17] Xue, J., & Su, B. (2017). Significant remote sensing vegetation indices: A review of developments and applications. Journal of sensors, 2017.
[18] Syafiqah, N. F., Shariff, A. R., & Che'Ya, N. N. (October 2019). GIS and UAV Aerial Imaging Applications for Durian Plantation Management. 40th Asian Conference on Remote Sensing, ACRS.
[19] Qiu, Z., Chu, X., Calvo-Ramirez, C., Briso, C., & Yin, X. (2017). Low altitude UAV air-to-ground channel measurement and modeling in semiurban environments. Wireless Communications and Mobile Computing, 2017.
[20] Eng, L. S., Ismail, R., Hashim, W., Mohamed, R. R., & Baharum, A. (2018). Vegetation monitoring using UAV: A preliminary study. International Journal of Engineering and Technology(UAE), 7(4), 223–
227.
[21] Bassine, F. Z., Errami, A., & Khaldoun, M. (2019, June). Vegetation Recognition Based on UAV Image Color Index. In 2019 IEEE International Conference on Environment and Electrical Engineering and 2019 IEEE Industrial and Commercial Power Systems Europe (EEEIC/I&CPS Europe) (pp. 1-4). IEEE
[22] Mallmann, C. L., Zaninni, A. F., & Pereira Filho, W. (2020, March). Vegetation Index Based In Unmanned Aerial Vehicle (Uav) To Improve The Management Of Invasive Plants In Protected Areas, Southern Brazil. In 2020 IEEE Latin American GRSS & ISPRS Remote Sensing Conference (LAGIRS)
4322 (pp. 66-69). IEEE.
[23] Bendig, J., Yu, K., Aasen, H., Bolten, A., Bennertz, S., Broscheit, J., ... & Bareth, G. (2015).
Combining UAV-based plant height from crop surface models, visible, and near infrared vegetation indices for biomass monitoring in barley. International Journal of Applied Earth Observation and Geoinformation, 39, 79-87.
[24] Cermakova, I., Komarkova, J., & Sedlak, P. (2019, June). Calculation of visible spectral indices from UAV-based data: Small water bodies monitoring. In 2019 14th Iberian Conference on Information Systems and Technologies (CISTI) (pp. 1-5). IEEE.
[25] Zhang, D., Mansaray, L. R., Jin, H., Sun, H., Kuang, Z., & Huang, J. (2018). A universal estimation model of fractional vegetation cover for different crops based on time series digital photographs. Computers and Electronics in Agriculture, 151, 93-103.
[26] Zhang, D., Mansaray, L. R., Jin, H., Sun, H., Kuang, Z., & Huang, J. (2018). A universal estimation model of fractional vegetation cover for different crops based on time series digital photographs. Computers and Electronics in Agriculture, 151, 93-103.
[27] Paramasivam, C. R., & Venkatramanan, S. (2019). An introduction to various spatial analysis techniques. GIS and Geostatistical Techniques for Groundwater Science, 23-30.
[28] Hyun, C. U., & Kim, H. C. (2018). Mosaicking very-high-resolution helicopter-borne images acquired over drifting arctic sea ice using COTS sensors.
[29] Kuželka, K., & Surový, P. (2018). Mapping forest structure using UAS inside flight capabilities.
Sensors, 18(7), 2245.
[30] Mateen, A., & Zhu, Q. (2019). Weed detection in wheat crop using UAV for precision agriculture.
Pakistan Journal of Agricultural Sciences, 56(3).
[31] Montilla, Ricardo, Guillermo Montilla, Egilda Perez, Luigi Frassato, and Cesar Seijas. "Precision agriculture for rice crops with an emphasis in low health index areas." Revista Facultad Nacional de Agronomía Medellín 74, no. 1 (2021): 9383-9393.
[32] Xue, J., & Su, B. (2017). Significant remote sensing vegetation indices: A review of developments and applications. Journal of sensors, 2017.
[33] Mateen, A., & Zhu, Q. (2019). Weed detection in wheat crop using UAV for precision agriculture.
Pakistan Journal of Agricultural Sciences, 56(3).
[34] Bouajaj, A., Bahi, L., Ouadif, L., & Awa, M. (2016). Slope stability analysis using GIS. The International Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences, 42, 151.
[35] Balasubramanian, A. (2017). Digital elevation model (DEM) in GIS. University of Mysore.
View publication stats View publication stats
