This AI project arose from the need of pineapple farmers to know how many fruits are in their fields. Until now, these farmers would take photos with their drones and send them to companies where pineapples were manually counted.
The objective of this platform is to eliminate any possible human error in pineapple counting through artificial intelligence, and to provide a report and different images within approximately 2 hours.
iHawk is a platform that allows image processing through artificial intelligence. It provides users with the ability to manage their images autonomously, obtaining processing results in just a few minutes.
The platform's potential lies in the development of an artificial intelligence capable of counting pineapples from an image of plantation fields. In addition, we offer the added value of a user-friendly portal, allowing users to obtain their results in just a few hours without the need for third-party assistance or technical knowledge.
Eliminating human error in counting
Image processing using AI
Creating an intuitive and simple portal
Once the images are obtained, our AI system will work by analysing them, generating three different images (terraces, plants, and density) and other data. Once the process is completed, the user can consult the information whenever they want and download the generated report.
We can talk about three differentiated parts in the development of the platform:
The web portal developed with the latest Microsoft technologies and together with a SQL database, all hosted on Azure. This portal is used by the user to upload their images and be able to consult those that have already been processed.
The image container, both those uploaded by the user and those generated by the system itself. These images are hosted on a specific Azure file system, Azure Storage File Shares.
An Azure virtual machine where the AI process developed for pineapple counting is hosted. Additionally, a queue management process is also hosted here, which is responsible for managing user requests according to the order of arrival and subsequently updating the processed image with the results.
The development of this solution has enabled a user to process the images of his crops quickly, safely and autonomously. In approximately 2 hours they can consult the results supported by the images of terraces, plant and density generated by the system itself.
Finally, a platform has been created in which any user can receive images and reports on the counting of their products and other relevant data tailored to their needs thanks to its artificial intelligence.