This week, the award-winning works of the 22nd edition of the Salão de Pesquisa SETREM (SAPS) were announced. In this edition, four articles were presented orally by the scientific initiation students: Anthony Vanzan, Gabriel Fim, Matheus Sausen and Greice Welter. The presentations were awarded 1st, 2nd and 3rd place in the Computing area and 2nd place in the Agronomy area. These are some of the results achieved in the AgroComputação project. The initiative to publish the articles came from the idea of sharing the research findings and publicizing the work that has been done on this project.

Next, the winning students tell how was that experience to present their works at SAPS.

Anthony Vanzan, student of Computer Engineering course of SETREM: 

“It was a new experience, since I had never presented at SAPS, the experience provided me learning and knowledge. The work presented was entitled “DEEP LEARNING ALGORITHM FOR CLASSIFICATION OF CROP AREAS WITH UAVs”, the objective of the work was to create a deep learning model for classification and demarcation of maize crops through aerial images taken by UAVs.The presentation of the algorithm and discussion of the results earned the work the 1st place in the oral category of SAPS 2020 in computing area. The work developed will be used later to carry out the calculation of planting density and estimate of maize crop production. “

Gabriel Fim, student of Information Systems course of SETREM:

“It was a new experience, I had already presented a work in a previous edition of SAPS, but this time because it is a completely online edition, the experience was very different from the previous one. The work presented was called “DEVELOPMENT OF AN AUTOMATIC PRE-PROCESSING ALGORITHM OF IMAGES TAKEN WITH UAVs”, aiming to present the algorithm developed within the AgroComputação project to perform the preprocessing of the images taken with UAVs and automated creation of the training dataset for developed of neural networks. In this way, I believe that the algorithm developed and presented in this work can serve as a basis for future projects that wish to carry out similar activities and projects, since during the development of the algorithm we were unable to find works that had a similar objective of performing the pre-processing of images in order to create more automated datasets. “

Matheus Sausen, student of Agronomy course of SETREM:

“It was a new and important experience both personally and professionally for the future, as I had never presented it at SAPS or at other events. The presentation was called “METHODOLOGY FOR CAPTURE OF IMAGES WITH UAV FOR MAIZE CROPS“, with the objective of defining an image capture methodology with UAV, to obtain metrics in the maize crops (Zea mays), because there were no standard methodologies to be followed in the literature to obtain metrics for that crop. Thus, I believe that this work has much to contribute to other researchers or technicians in the area, who wish to apply or carry out new research on the use of UAVs in agriculture. Agriculture in recent decades has undergone countless transformations, with the objective of improving the production system and also increasing the levels production, having a direct link with technological advancement.

Greice Welter, student of Computer Engineering course of SETREM:

“It was a new experience that added knowledge in several aspects, both personal and professional, since I had never presented it at SAPS before. The work presented was “EVALUATION OF FRAMEWORKS TENSORFLOW, PYTORCH AND KERAS FOR DEEP LEARNING“, the aim was evaluating frameworks for the development of artificial intelligence in the AgroComputação project, where we sought to evaluate the best framework for the detection of tassel and maize plants with images taken by UAVs. We believe that this study can help others in choosing the most efficient and appropriate framework for object detection. “

Check out all the award-winning works at SAPS 2020 here.