GreenGuard: Guard Your Crops, Grow Your Business

Inspiration

One of our team members has a family involved in the agricultural industry, and for years they have observed that one of the biggest issues in harvest optimization is crop disease. At times, entire batches of long-nurtured products are wasted due to late disease detection. After further research we were able to gather that this isn't an isolated case. According to the Mexican government, around ten thousand tons of crops are lost annually. Often, greenhouse-based producers focus on maximizing output, which can lead to neglecting waste management. Thus, we aimed to find a solution to prevent plant disease and promote a quality-based system that enhances business in the agricultural sector.

What it does

What GreenGuard does is serve as a medium for both greenhouse health monitoring and business networking. It is a user-friendly app powered by advanced software, including an AI model optimized through OpenVino, which provides a comprehensive and personalized dashboard for monitoring greenhouse health. The system incorporates historical data and offers evidence-based certification to help connect producers with buyers, all while ensuring product excellence.

How we built it

The development of GreenGuard involved a robust set of tools and technologies to ensure a seamless, scalable, and efficient application. For machine learning, TensorFlow served as the foundation for building our models, while OpenVino optimized these models for enhanced performance on various hardware architectures. Additionally, we integrated Softtek's Frida, which enabled us to efficiently monitor and manage the app’s real-time processes. On the backend, we leveraged Django to create a solid web framework that supports our application’s data handling and API functionality. To ensure secure and scalable deployment, we utilized Google Cloud Platform (GCP) for cloud hosting and infrastructure, along with Docker for containerization, simplifying deployment and ensuring consistency across different environments. For image processing and manipulation, Pillow was used, providing tools to analyze crop images for health assessment. Together, these technologies allowed us to create a powerful and user-friendly solution for greenhouse health monitoring and business networking.

Challenges we ran into

During the development of GreenGuard we ran into multiple challenges in different areas of expertise. For one, adapting our deep learning models to diverse dataset layouts proved to be, if not difficult, tedious. Plus, training the different models and integrating OpenVino took a significant amount of time. Working with the cloud also required extensive effort, as we had to familiarize ourselves with a large amount of documentation. However, the biggest obstacle we faced was integrating the various components into one coherent piece of software. While each part of GreenGuard's development presented a challege of its own, combining them presented the real difficulty.

Accomplishments that we´re proud of

We believe the final product of our efforts during this Hackathon is highly comprehensive, offering a range of valuable services to users while integrating multiple areas of software development. Despite the ambitious scope of our project, we successfully combined every aspect seamlessly into a cohesive whole.

What we learned

Throughout the project's development, we had the opportunity to work with new technologies, such as Frida and OpenVino. This experience taught us how to effectively integrate these tools into our code to create a better product. We also gained valuable insights into working with multiple branches and organizing them efficiently to ensure smooth project progress.

What´s next for Git Push It Real Good

As a team, we are very pleased with our project idea and its execution, and we plan to continue developing it further. We aim to add a variety of new features while refining the existing ones. By creating more powerful deep learning models with near-perfect accuracy, we can expand the scope of our project. We truly believe that GreenGuard has the potential to make a significant impact on optimal crop harvesting and the agricultural sector as a whole.

Built With

• cloudstorage
• css
• django
• docker
• frida
• googlecloudplatform
• html
• javascript
• openvino
• pillow
• postgresql
• python
• sqlite
• tensorflow