Satellite Imagery, GIS and AI Boost Date Palm Production in Oman

The Sultanate of Oman has initiated a ‘million date trees plantation project’ and is using Satellite Imagery, Geographic Information Systems (GIS) and Artificial Intelligence (AI) services provided by Hardcastle GIS Solutions, a global geospatial technology provider company based in India, to increase the production of dates in Oman.

Date palms are the primary agricultural crop in Oman. The country produces more than 250,000 metric ton of dates per annum, but is still ranked only 8^th in terms of production in the world. Oman wants to transform the way its date plantations approach the cultivation process in order to increase productivity, utilize resources correctly and efficiently, and improve quality. Therefore, it has initiated a ‘million date trees plantation project’ that will utilize Satellite Imagery, GIS and AI technology services offered by Hardcastle GIS Solutions for this purpose.

Workflow

Hardcastle has mapped the study area using 0.5m Satellite Imagery, which is then processed and used to create GIS layers for analysis.
Hardcastle has also conducted a physical ground survey using GPS and collected more than 20+ attributes for each tree, tagging the trees using a Physical Specialized All-Weather Printed QR Code.

Using a GIS-based survey that comprises attributes essential for the production and quality of the date trees, Hardcastle has captured the GPS location of every tree and marked it with allocated Physical Printed Specialized All-weather Digital codes, and geo-tagged the trees.

Hardcastle has used its software to compile details of the farms, cultivators, tree codes, and tree numbers. Details of the number of fronds on trees, their average length, and average girth of the trunks have also been captured and stored. Each tree has a unique number and has been photographed.

All the data is stored in a cloud-based ERP system, which has been integrated with Hardcastle’s Open Source GIS Web Application.
From pesticides to water level to temperature, all the data are collected in ERP systems and each date tree is given different levels of inputs that are tracked using sensors, etc.
Drones, tractors, CCTV cameras, and other special equipment are used in the farms to capture real-time plantation site information. The inputs from these systems are processed using Artificial Intelligence/Machine Learning (AI/ML) technologies, which alert the authorities for any insect damage or any critical parameters that can impact the quality of the dates. These AI-processed data inputs are then automatically sent to a GIS system from where the control team takes the necessary actions.
Once the dates are procured, the quality assurance team checks and matches the most premium quality dates, referencing them to the particular tree that has given the best output. These data parameters then act as input parameters for the next cultivation cycle, and so on.

Benefits and outcomes of the project with specific focus on the value and utility of geospatial solutions

Access to reliable and up-to-date crop production information is essential in the agriculture sector today. Whether one is managing a large agro-holding or running a farm, GIS offers satellite-based crop management products that provide detailed insights into the state of the crops and assist in decision-making throughout the season.

Tree data analyses and real-time insights on tree requirements concerning water, fertilizer, health, and other attributes provide agility to the system, which in return helps in making decisions actionable in real-time.
Total production has increased by 40% and the quality of the dates has also improved.

Challenges and learnings from the project

Conducting a physical survey at a desert farm, with temperatures above 35°-50° Celsius, proved very challenging. In the desert, there isn’t much vegetation for the survey team to rest under. Also, desert snakes often take shelter beneath the few trees and the team encountered this situation at times. The team had to adjust to local timings to start the GIS surveys as early as 6 am and end before 10 am. The surveys were restarted post 3 pm to avoid the heat and sun as much as possible.
Capturing 20+ attributes of trees isn’t an easy job; counting girth, the number of offshoots, roots, and other parameters requires an understanding of date tree agriculture at a detailed level.

The use of AI/ML training datasets was challenging for the surveyors. For example, they could not find enough sample images to identify the red palm weevil. It was good learning for them, however, to use a Drone platform to manage and prevent agricultural pests, particularly the red palm weevil.

This pilot project has helped define an enterprise-wide GIS policy. The project will now be executed at the enterprise level across the country. It will help establish standards and showcase the benefits of GIS for the agriculture sector.