AN AI BASED ADVANCED AG-DRONE SYSTEM FOR LOCAL FARMERS IN SC

• Sponsoring Institution: National Institute of Food and Agriculture
• Project Status: COMPLETE
• Funding Source: EVANS-ALLEN
• Reporting Frequency: Annual
• Accession No.: 1022704
• Project Start Date: May 16, 2020
• Project End Date: May 15, 2023
• Program Code: [(N/A)]- (N/A)

Recipient Organization
SOUTH CAROLINA STATE UNIVERSITY
(N/A)
ORANGEBURG,SC 29117

Performing Department
Engineering Technologies

Non Technical Summary
According to UN Food and Agriculture Organization, by 2050 the population will increase by 2 billion and only 4% additional land will come under cultivation. And by USDA's National Agricultural Statistics Service (NASS) and ERS, only 70-80% of larger farms are adopting existing agri-technologies for both high yield and profit. No participation from local and small farms due to the fact that the existing agri-technology doesn't provide an economical integrated solution to local or small farmers to get benefits on data storage, technology, inputs, and market. Thus, this research aims to provide a low cost integrated advanced Ag-drone based agrisystem to local farmers in SC. This research project will use different equipment such as drones for taking aerial images of crops for health assessment, automatic irrigation control, ultrasound sensor for reading soil water level and plant root monitoring, AI modules for data analysis and access to internet cloud for data storage, Website/blog for smart farming related knowledge sharing. The final agrisystem will implement all the above-mentioned objectives to be available to local farmers for testing and use. A low cost and affordable integrated advanced agrisystem will be achieved through the above-mentioned methods which will impact the local and small farming method by increasing yield, quality product, and better profit. Also, other benefits are achieved such as accessing and sharing agriculture related knowledge to the related communities in the region, nation.

Animal Health Component

25%

Research Effort Categories

Basic

75%

Applied

25%

Developmental

0%

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
404	7210	2080	100%

Knowledge Area
404 - Instrumentation and Control Systems;

Subject Of Investigation
7210 - Remote sensing equipment and technology;

Field Of Science
2080 - Mathematics and computer sciences;

Keywords

precision farming

ai or machine learning

ultrasound imaging

automatic irrigation

plant health monitoring

Goals / Objectives
The major goal of this research is to develop an economical agrisystem that will achieve increased product yield and quality, automatic irrigation control, data analysis, intelligence sharing with other local farmers, making research data available to other research community.1. Development of noninvasive ultrasound sensor2. Drone controlled automatic irrigation system3. Implement Drone mounted computing4. Monitoring plant starting from root to leaf to ensure quality5. Implement cloud storage and web data hosting

Project Methods
The project aimed to develop a drone based agrisystem where a unique noninvasive ultrasound imaging will be used to measure soil water level rather using a traditional sensor and to monitor the plant root health through ultrasound image analysis. Also, drone mounted computing will be implemented which is another unique method to provide a real-time data analysis. During the initial project period, the system will be analyzed and evaluated in the laboratory environment. Once the system completely developed, it will be tested on the actual farms. The project efforts will increase knowledge on the application of ultrasound imaging in precision farming technology, knowledge on safe operation of drones, field and crop image analysis, over the year collected big-data analysis, drone controlled automatic irrigation on small/local farms, seminars on agrisystem for students. The final developed system will be evaluated on the local farms in Orangeburg, Bamberg, St. Mathews, Bowman for its impact in increasing yield, quality and profit, automatic irrigation, better monitoring, Knowledge sharing and privacy. The project results will be published in conferences and journals and also, made available to public through website.

Progress 10/01/20 to 09/30/21

Outputs
Target Audience:Local farmers and small farmers in Orangeburg, Bamberg, Bowman, and St. Matthews. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?This project created one-on-one work with student mentoring activities. The principal investigator (PI) had provided training on python to the new undergraduate research students on a need basis, this included steps to solve research problems, implement research algorithm using python programming. Also, provided sessions on how to read and write technical research papers. How have the results been disseminated to communities of interest?Two conference papers are submitted: SoutheastCon 2022 (manuscript Under review) and ASEE (abstract accepted). Also, abstracts are accepted for Student and PI presentations at ARD symposium, Atlanta, April 2-6, 2022. What do you plan to do during the next reporting period to accomplish the goals?To accomplish the project goals and objectives, our team has the following plans to do for the next reporting period. 1) We will address the challenge of automatic drone operation. 2) Drone Security challenges.

Impacts
What was accomplished under these goals? This project will primarily help the small and local farmers in the state of South Carolina, especially nearby local farmers in Orangeburg, Bamberg, Bowman, and St. Matthews counties. Under the progression of the project three computer science undergraduate students and one engineering student have gained valuable research experience that is directly related to this project. Also, our team is anticipating for the greater awareness of farming technology usage within the local farmers. For this reporting period, the following activities are accomplished related to the goal-"Implementing Drone Controlled Automated Irrigation System". 1) Major activities completed AI/ML Implementation: The Artificial Intelligence software components such as Convolutional Neural Network (CCN) and Deep Learning (DL) are implemented using Python programming language. Soil Moisture Determination using AI applied to Drone Images - The rapid adoption of Artificial Intelligence (AI) and drones see many precision farming applications such as disease detection from the image, identification of crop-readiness, farming field management, monitoring crop health, soil profile and active irrigation automation. The proposed method is an AI technique - deep learning based noninvasive technique applied to drone captured images of soil to determine soil moisture supported along with ultrasonic soil profile. DeepTrac: Applying Artificial Intelligence in Plant Disease Detection - Most critical and challenging problem in farming is the early detection of plant diseases. In this research, an AI-based automated plant disease detection method using FPN with Faster R-CNN architecture is proposed. 2) Data collected - The collected data are the soil images and plant leaf images from that are captured by the drone. Further, these captured image data are annotated using a computer vision (CVAT) tool. 3) Two conference papers are submitted (Under review). SoutheastCon 2022 and ASEE conferences. 4) Abstracts are accepted for Student and PI presentations at ARD symposium, Atlanta, April 2-6, 2022.

Publications

• Type: Conference Papers and Presentations Status: Under Review Year Published: 2022 Citation: Biswajit Biswal and Simien Chestnut. DeepTrac: Applying Artificial Intelligence in Plant Disease Detection submitted to SoutheastCon 2022 conference will be held in Mobile, Alabama from MARCH 31st to APRIL 3rd, 2022 (Virtual).
• Type: Conference Papers and Presentations Status: Accepted Year Published: 2022 Citation: Biswajit Biswal and Terrence Duncan. Soil Moisture Determination using AI applied to Drone Images at ASEE Annual Conference and Exposition, Minneapolis, Minnesota, June 26-29, 2022.
• Type: Conference Papers and Presentations Status: Submitted Year Published: 2022 Citation: Biswajit Biswal. An AI Based Advanced Ag-Drone System for Local Farmers in S.C. Department of Computer Science and Mathematics, South Carolina State University, Orangeburg, SC 29117. Submitted at ARD symposium, Atlanta, April 2-6, 2022.
• Type: Conference Papers and Presentations Status: Submitted Year Published: 2022 Citation: Simien Chestnut and Biswajit Biswal. Using Artificial Intelligence with Python to Detect Plant Disease. Department of Computer Science and Mathematics, South Carolina State University, Orangeburg, SC 29117.Submitted at the ARD symposium, Atlanta, April 2-6, 2022.

Progress 05/16/20 to 09/30/20

Outputs
Target Audience:Local farmers and small farmers in Orangeburg, Bamberg, Bowman, and St. Matthews. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?This project created one-on-one work with a mentor training activity. The principal investigator (PI) had provided training on python to the undergraduate research student on a weekly basis, this included steps to solve research problems, implement research algorithm using python programming. So far, the project has formed a research team which includes undergraduate research students and the principal investigator. A graduating research undergraduate student has gained valuable research experience in the following areas: 1) Learned a new programming skill. 2) Implemented some research algorithms. 2) Developed python software packages. 3) Tested and validated the developed software with data. Also, our research team is anticipating to setup an experiment testbed, equipment, collect data and produce results. How have the results been disseminated to communities of interest?Since this project has recently started and due to the COVID-19, there are no outreach activities that have been undertaken at this moment. However, our team is trying all the necessary steps to be able to communicate with the local farmers. What do you plan to do during the next reporting period to accomplish the goals?To accomplish the project goals and objectives, our team has the following plans to do for the next reporting period. 1) Develop software and hardware to monitor plant health such as plant disease, growth, etc. 2) Implementing Drone Controlled Automated Irrigation System - hardware implementation and software implementation for automatic irrigation system. 3) Software implementation of AI (Artificial Intelligence) models.

Impacts
What was accomplished under these goals? The major goal of this project is to develop a low-cost hi-tech agrisystem to investigate the logistics of an economical precision farming which increases product yield, quality, profit, and intelligence sharing with other local farmers. And the major objectives of this project are: 1) Preparing and Mounting Ultrasound Sensor in Drone; 2) Implementing Drone Controlled Automated Irrigation System; 3) Implementing Automatic Drone-Hop; and 4) Implementing Drone mounted Computing. This project will primarily help the small and local farmers in the state of South Carolina, especially nearby local farmers in Orangeburg, Bamberg, Bowman, and St. Matthews counties. Under the progression of the project one graduating computer science undergraduate student has gained valuable research experience that is directly related to this project. Also, our team is anticipating for the greater awareness of farming technology usage within the local farmers. For this reporting period, the following activities are accomplished related to the goal-1 "Preparing and Mounting Ultrasound Sensor in Drone". 1) Major activities completed Software: Wrote python programs to make the ultrasound sensor to capture 1D images and analyze those images. Implemented a random signal generator using python that acts as input to the butterworth bandpass filter. Implemented the butterworth bandpass filter equation using python. Python module that calculates the distance to an object is based on the received signals reflected from an object and converts the received signal strength to numeric measurement value. Wrote a python program for the signal envelope detection via a Hilbert transformation. Integrated developed python program modules that take images as input, filter out noises and reconstruct a final image. Implemented a python program to convert polar coordinates to Cartesian coordinates. Wrote a python program to log-compress and separate the signal envelopes from background noise. Hardware: Ultrasound sensor is developed using 3 ultrasonic sensors, servo-motors to capture signals out of which 1D images are constructed through the image construction software. 2) Data collected - The collected data are the reflected signals from an object that are captured by the sensors. Further, these captured data are processed through the image construction software to produce 1D image. 3) Summary statistics and discussion of results: The team is working on to produce results. 4) Key outcomes or other accomplishments realized: Python software packages and hardware sensors.

Publications