CoFly-GUI is a novel, low-cost, user-friendly Precision Agriculture platform that attempts to alleviate the drawbacks of limited battery life by carefully designing missions tailored to each field's specific, time-changing characteristics. The proposed platform aims at providing the operator/farmer with a cost-effective and end-to-end integrated system that is able to accomplish autonomous tasks that were used to completed manually (e.g. crop growth monitoring). Besides, it will play a supportive role in decision-making objectives (e.g. possible weed infestations) enhancing crop yield management decisions by assisting non-expert users in proper manipulations. In specific, the proposed system is capable of:
- UAV built-in flight planning for automated crop monitoring (Waypoint-Trajectory-Planning module)
- Integrated image processing functionalities for orthomosaic extraction and vegetation health estimation (Vegetation-Indices module)
- Automatically extracts possible problematic areas of the field (Problematic-Areas-Detection module) and subsequently designs a follow-up UAV mission to acquire extra information on these regions
- UAV built-in site-specific mission based on the extracted knowledge and combined with incorporated deep learning model (Weed-Detection module) trained on the CoFly-WeedDB dataset for weed detection
- Timeline view for storing information according to the year's harvest
All the above functionalities are enclosed into an open-source, end-to-end platform, named Cognitional Operations of micro Flying vehicles (CoFly). The effectiveness of the proposed system was tested and validated with extensive experimentation in agricultural fields with cotton in Larissa, Greece during two different crop sessions. In Figure 1, we present the architecture of CoFly.
Figure 1. CoFly’s high-level architecture.
- Clone this repo
- Download & install Node.js
- Open terminal on ~REPO_PATH
- Install all necessary dependecies
npm install
- Run CoFly-GUI
npm start
The developed platform offers capabilities that fit each field's characteristics, such as no-fly zones inside the area of interest and automatic UAV-based weed detection. In this section, we present the operations and data postprocessing capabilities of this application.
When CoFly-GUI is installed, the above screen is the main menu. The main menu has 3 options:
- Load project (see an existing project)
- Create project (create a new project)
- Import project (import a project from a hardware system to another).
The blue polygon refers to the field of interest for further examination. The bounds of field of interest are set manually by the user.
After setting the scanning area, we export the UAV mission path with the Waypoint-Trajectory-Planning module. We demonstrate the extracted UAV path planning with and without obstacles (red color). The obstacles are set manually by the user. Also, the end-user has the ability of knowing the position of UAV 
.
When the UAV mission completes, the built-in stitching process starts automatically using the visual footage as captured by UAV during its mission. The result of image stitching process is the orthomosaic of the under study field. Also, some update messages appear in order to inform the end-user about the progress of the pipeline.
The result of image stitching process is further processed in order to calculate the vegetation indices (VIs) with the Vegetation-Indices module. Each VIs represents the actual reflectance of the examined field’s vegetation in different color bands and thus, it can reflect different measures of crop health.
Based on VIs maps, we identify the problematic areas of the under study field in terms of vegetation health using Problematic-Areas-Detection module. For every detected area, the center of mass is calculated leading to a set of points, considered as points of interest. The end-user has the option to see the UAV nearest captured image and remove them. Each point of interest is annotated with 
.
Hot-Poin Mission focuses on the detected points of interest for each VI. Again, the UAV path planning is extracted by the Waypoint-Trajectory-Planning module. During this mission, visual footage from lower altitudes and different camera angles is captured aiming to visually detect crop health deterioration sources, such as weed plants, in order to enable site-specific treatment.
The developed operation of CoFly-GUI is capable of semantically segmenting weed instances depicted on input RGB images as captured during the Site-Specific Mission and thus, provides accurate information regarding the location of detected weeds (Weed-Detection module). In this way, the end-user has the ability to schedule counteractions on the spot.
A cropping calendar is a schedule of the crop growing season from the fallow period and land preparation, to crop establishment and maintenance, to harvest and storage. Using this allows better planning of all farm activities and the cost of production.
All the captured UAV images during the coverage mission are availabe in PHOTO GALLERY option. These images are seperated per date and time of UAV mission.
CoFly-GUI is an intuitive graphical user interface including higher level user commands and proper data representations depicting the results of the UAV’s operation. The final product is capable to accomplish tasks previously completed manually (e.g. crop growth monitoring) and play a supportive role in decision making objectives (e.g. possible insect infestations).
(not published yet)
This research has been financed by the European Regional Development Fund of the European Union and Greek national funds through the Operational Program Competitiveness, Entrepreneurship and Innovation, under the call RESEARCH - CREATE - INNOVATE (T1EDK-00636).