Using AI and Drones to Create Precision Agricultural Systems for Commercial Farming

Disclaimer: Aya Data holds to the confidentiality agreements made with our clients. We shall not disclose any confidential information regarding clients and/or projects, pursuant to the individual agreements made. Consequently, client and project names and any identifiable information may be kept anonymous in our Case Studies.

Overview

Managing large-scale commercial farms necessitates a lot of progress tracking and repetitive tasks, in addition to being manually overwhelming. However, crop management systems powered by drone-fed computer vision models are a solution to the traditional system of manual data collection, while improving data accuracy and providing valuable insights into farming operations.

Data Science and AI

Image segmentation

Industry

AgTech

Headquarters

London, UK

Company Size

250+

drone flying

Challenge

The Client is the owner of a large-scale, 6,000-ha palm plantation in Ghana. They were having trouble accurately monitoring the plantation and accessing digitized information in real time. Manual data collection and monitoring had led to inaccurate plant counts and a lack of data-driven insights, in addition to high labor and resource costs.

The Client decided that their monitoring operations could be optimized by implementing machine learning (ML) solutions to streamline their internal processes.

Solution

Aya Data was hired to capture RGB and LiDAR data across the plantation to implement the solution. We collaborated with drone partners for the data collection aspect of the project. This data was processed and provided to our in-house data science team as GeoTIFF files.

Utilizing a collection of aerial images, Aya Data’s annotation specilists labeled the datasets using bounding boxes and keypoints for the palm counting model. The data science team used the labeled data to create and train the palm counting model, which had a 98% accuracy rate regarding tree count and location.

The solution was deployed to the client using the ArcGIS third-party dashboard, allowing for visualization of plant count and other relevant parameters of the farm using the drone data.

Contact us

Results

The overall effect of Aya Data’s involvement and the implementation of the crop management dashboard was a significant improvement to the Client’s farming operations:

  1. The number of trees on the plantation could be accurately counted, which served as a reliable baseline for various operational decisions;
  2. The dashboard provided insight into plant health, which enabled the client to focus their resources only on the areas that required them, ultimately lowering operational costs;
  3. The automatization of the monitoring process reduced the time and effort necessary to manually monitor the entire plantation.

Future Phases

The second phase of the project involves integrating the crop management dashboard with a Business Intelligence (BI) system. This integration will provide the company with comprehensive insights into their daily, weekly, and quarterly inspection exercises, streamlining all their farm operations and data into a centralized dashboard. This advancement will enhance data-driven decision making for the management team, further optimizing their farming processes.