The paper entitled Simplifying the Interpretation of Soil Analysis Reports with Deep Learning in Cloud, recently published in Argentina-Brazil Electronic Journal of Information and Communication Technologies (Revista Eletrônica Argentina-Brasil de Tecnologias da Informação e da Comunicação, or simply REABTIC), is the result of the undergraduate thesis from graduates Information Systems students Alisson Allebrandt and Diego Schmidt with advisorship from Prof. Dalvan Griebler. The application which was created is a first step to help farmers to interpret the soil analysis and understand what correction needs to be applied to the soil depending on the results of the analysis.
The contributions of this research are summarized below:
- Proposal for a software architecture to assist in the interpretation of soil analysis reports.
- Implementation and evaluation of two cloud services to character recognition based on Deep Learning.
- App to assist producers and agronomists in the process of analysis and interpretation of soil analysis reports.
Alisson and Diego discuss the experience of developing this research.
The development of this work provided knowledge and new experiences both in the technical area and in the agriculture area through direct contact with agronomists engineers and with the soil analysis laboratory of Setrem, who helped us and opened the doors for us to develop our idea. During the development of the work, it was possible to approach artificial intelligence tools and OCR interpreters, such as the Google Vision API and Tesseract OCR, which are widely used in the industry today. In addition, we applied these technologies in cloud environments to perform interpretations of images of soil analysis reports. The knowledge acquired during the work was great and brought us new insights into the area of agriculture and especially in cloud computing technology, which drove me to start a postgraduate in Cloud Computing in Setrem.
Alisson
The development of this research provided new experiences and contacts with the agriculture area, more specifically in the soil area, because we talked with professionals and sought to understand more about the problems they faced. With the use of cloud computing, OCR techniques, and artificial intelligence, supported with tools known in the industry, such as Tesseract OCR and Google Vision, we were able to streamline the process of interpreting soil analysis reports. All the experiences we went through during the development of this work contributed both for personal and professional evolution, since we acquired new knowledge and improved our communication skills.
Diego
Check below the abstract of the paper. The full text is available following this link.
Abstract: Since the soil impacts directly on agricultural productivity, its conservation through the correct application of nutrients and fertilization is of paramount importance. In this work, we propose a software architecture and a mobile application capable of assisting farmers and agronomists in interpreting soil analyses generated from laboratories. The software architecture was designed for cloud environments and the mobile application is the interface for capturing and presenting data. Initially, it was necessary to create a database with different image types and configurations. All images from the dataset were treated to eliminate noise (such as brightness, shadows and distortions) and used to evaluate two Deep Learning solutions (Google Vision and Tesseract OCR), where Tesseract OCR proved to be more accurate using the same images. In addition to offering the mobile application, which is the first step, the research carried out reveals several technological deficiencies and opportunities for innovations in the field of soil science.