Study of the vegetation of spring crops in the region of South Dobrudzha in 2020

The present study observes the development of spring crops maize and sunflower typical of the Southern Dobrudzha region. A distinct methodology for remote monitoring through the use of a small remotely piloted aircraft has been developed for the purposes of the present paper and applied accordingly. Two types of video cameras were used: for the visible range of the RGB light reflected by the plants and for the NearRed reflection close to the red light. The obtained results are presented in both tabular and graphical form and inferred, finally, in the paper are some principal conclusions about the condition of the crops under consideration .


Introduction
Indices based on remote sensing to quantify the state of vegetation, namely the combination of recorded reflections of visible and infrared bands, provide a unique opportunity to track the characteristics of areas where there is vegetation -biomass in a state of growth and covering areas with leaf mass. These indices serve as a basis for assessing vegetation, (Wu, 2015). The relationship between the chlorophyll content of the plants and the light reflected by them allows for a more objective evaluation of the condition of the crop in terms of diseases and abnormalities. Some of the most widely used indices are: NDVI -Normalized Difference Vegetation Index, (Rouse, 1973), EVI -Enhanced Vegetation Index, (Huete, 2002), TCI -Temperature Condition Index, (Kogan, 1995), NDWI -Normalized Difference Water Index, (Gebrehiwot, 2016), VCI -Vegetation Condition Index and SPI -Standardized Precipitation Index, (Mckee, 1993). Indices based on remote sensing data are widely used mainly in the assessment of the condition of crops, as they give an accurate idea of the agricultural conditions for the observed area and facilitate the planning of the necessary activities for the needs of precision agriculture. Another type of reflective index based on infrared light is the LAI -Leaf Area Index considered by, (Asrar, 1984). It gives the relationship between the spectrum of infrared and red light reflected from wheat and soil.
An in-depth assessment of the vulnerabilities of cultivated plants caused by complex environmental and anthropogenic factors is constantly required due to climate change. Drought as a natural phenomenon seems to have many multidimensional effects on agriculture, (Singh, 2014). Various drought assessment indices have been developed from climate and satellite data and, provided on that account, are the respective algorithms for estimating the vegetation fractions in wheat for a wide range of soil reflections, (Gitelson, 2002). Assessments of cultivated crop and crop rotation are reviewed by, (David, 2014). The possibilities and disadvantages of using different ways of measuring reflected light in passive observation methods are further explored as given by (Meroni, 2009).
Despite the increased resolution of satellite monitoring, it is not accurate enough due to the fact that the size of one pixel is 5 to 10 m. Due account, moreover, should also be taken of their dependence on weather conditions . These data are useful for monitoring globally through the world . More detailed values are obtained by observing from a lower altitude, which can be achieved with aircraft or drones. Shooting from a low height gives an accuracy of centimeters and does not depend on the cloud cover of the ground. It can be done for and at a time that is important to obtain accurate data on the condition of crops.

Page | 123
The aim of the research is to compile and experiment a methodology for remote monitoring of agricultural spring crops of corn and sunflower. As a result, to collect data on the vegetation, respectively, their condition in different regions of South Dobrudzha. Create digitized maps of the reflected light from fields planted with spring crops in the near-infrared range of the sunlight reflected by the crops.
The task of the research is to capture certain sections of the corn and sunflower fields in different regions of South Dobrudzha by means of a remote-controlled aircraft equipped with RGB (Red, Green Blue) and RGN (Red, Green, NearRed) cameras. This digitized and geographically localized photographic material should be processed with specialized software products to calculate the respective vegetation indices, of which the NDVI index is the most widely used.

Material and Method
Two remote-controlled aircraft models DJI Mavic 2 and DJI Mavic Pro were used to conduct the study, which are controlled by a software application that is installed on a smartphone or tablet, (DJI GO 4. Retrieved from: https://www.dji.com/bg/downloads/djiapp/dji-go-4).
The RGB camera of the Mavic 2 drone is a model L1D-20c_10.3_5472X3648. It is equipped with a second special camera model: MAPIR Survey 3W_RGN_3.4_4000X3000. It has a promising lens and photosensor type: Sony Exmor R IMX117 12 MP (Bayer RGB) and outer GPS/GHSS unbox UBX-G7020-KT. In it, the blue color channel is modified to perceive the reflected solar radiation with a wavelength 850 nm. The technical characteristics of the photosensor of this camera are: width 4.044 mm, height 3.033 mm, pixel size 1.011 µm, main point X 2.022 mm, main point Y 1.5165 mm. Specialized software: (Mapir Calibrating Application. Retrieved from: https//www.mapir.camera/pages/calibrating-images-in-mapir-camera-control-application) and (Mapir Camera Control. Retrieved from: https//www.mapir.camera/collections/software).
The RGB camera of the Mavic Pro drone is model FC220_4.7_4000x3000. Sensor Dimensions: 6.367 [mm] x 4.775 [mm], matrix 1/2.3" CMOS, with effective pixels 12.35 МР; lens with viewing angle 78,8 0 ; focal length 20 mm; aperture f / 2.2; image distortion less than 1.5%; focal length of 0.5 m to infinity; ISO for photos from 100 -1600, for video from 100 to 3200, maximum image size 4000x3000; image format * .JPEG. Published: 2020-12-31 Figure 1 shows the places where the photographic material was collected. The approach in identifying it was to be relatively evenly distributed in the region of Southern Dobruja and to have massifs mainly of corn and sunflower, as in the study period (July 2020) they are still in the vegetation phase.
The methodology for surveying the areas sown with corn and sunflower includes the following activities: giving information about the location of the site: no airport nearby, long-distance pipelines, crowds, etc.; providing accurate information about the meteorological situation in the marked object for shooting. The most important parameters are: wind speed -should not be more than 8 m/s and cloud cover should be less than 30%. It is recommended that it is not low or fragmented, and that there is no possibility of precipitation and gusts of wind. Authorization should be obtained from the tenants of the agricultural land and from the aviation agency, as both drones fall into subcategory (IA) -A1. They have a take-off weight of about 900 g, (Regulation (EU) 2019/945 of the European Commission of 12 March 2019 and Regulation (EU) 2019/947). Before the flight of the aircraft, it is set for flight mode (Table 1), which includes: altitude -about 90 m, speed not more than 10 m/s is determined by the wind speed as well as the requirement to overlap the photos. Maximum distance from the place of departure not more than 1.5 km. The flight is limited, also in terms of duration, which depends on the electrical capacity of the drone batteries, its remote control and the NIR camera. The efficient and safe flight of the two drones: Mavic 2 Pro and Mavic Pro to return to the starting position is within 20 minutes with one battery charged.
The altitude of the flight affects the number of photos that the cameras take and the resolution of the images GSD (Ground Sampling Distance), (Pix4D_GSD_Calculator. Retrieved from: https//support.pix4d.com/hc/en-us/articles/202560249-TOOLS-GSD-calculator) and this is the minimum distance on the ground from which information is recorded in one pixel of the video sensor. When the flight altitude is greater, the resulting image covers a larger area, but the resolution decreases, i.e. the distance corresponding to one pixel increases. The optimal combination is at a height of 90 m the resolution is within 3 cm per pixel! The overlap of the images is also important for the subsequent processing of the photographic material. This overlap should be within 85% of the flight direction and 70% to the left and right, (Pix4Dmapper. Retrieved from: https//www.pix4d.com/download-software). These parameters are usually set in the software application used to plan the route of the drone. It is an assistance program, (Pix4DCapture. Retrieved from: https://www.pix4d.com/product/pix4dcapture) which is installed on the tablet or smartphone that controls the drone.

Page | 125
The obtained maps of the reflection after the processing of the photographic material represent digitized information, the processing of which, with appropriate techniques, allows the formulation of various quantitative criteria (vegetation indices) for assessing the condition of the captured crops. In the present study, two groups of three indices were selected, which are formed by the reflections obtained from both types of cameras. The name, purpose and formula for the respective indices are given in Table  2 for those formed with RGB camera data and in Table 3 for those formed with RGN camera data. The criteria by which these two triplets of vegetation indices are selected are: the frequency of usage and the provision of objective information about the condition of the substance chlorophyll in the plants, vital for their good "health".

3
Results and Discussion For the calculation of the vegetation indices it is important how the weight coefficients will be selected for the values of the reflection recorded on the respective channels. Information on this is given in Table 4. The values of the weights shown are selected on the basis of information from comparative previous observations.
The results for the calculated vegetation indices in numerical terms as a range of change from maximum to minimum value are recorded in Table 5 for the indices formed with data from the RGB cameras by dates, places and cultures.

Page | 126
The results for the calculated vegetation indices in numerical terms as a range of change from maximum to minimum value are reported in Table 6 for the indices formed with data from the RGN camera by: dates, places and cultures.    (Table 5 and Fig. 2 and 4) shows good condition of the observed crops. The most frequently used vegetation index NDVI for the observed maize (Table 6 and Fig. 3) is in the range from 0.6 (village of Izvorovo) to 0.84 (village of Vedrina), which shows that these crops are in very good "health", i.e. in very good condition in terms of vegetation and no presence of diseases. With the lowest possible limit 0.3 being registered in the village of Izvorovo, which is ascribed to the sharp changes in the meteorological conditions during the process of observation, namely-clouding, which strongly affected the parameters of the reflected light.