Agro M2

Awards & Nominations

Agro M2 has received the following awards and nominations. Way to go!

Global Nominee

Crop Health Assistant 4.0.

High-Level Project Summary

The Crop Assistant 4.0 assists the producer in decision making, this happens through an artificial intelligence that will be trained to identify when the producer must intervene on an insect population that is in the crop, and also helps even to identify which insect would be, based on the analysis of the leaf that this insect fed on. The assistant follows the line of the challenge that to develop an innovative way to advance the ability to detect insect life, in this case, the ability to detect insects considered pests, this would be done based on satellite data, which would later be passed on to a trained artificial intelligence. It's important, it helps producers make important decisions.

Link to Project "Demo"

https://youtu.be/xztYpDDtXRc

Link to Final Project

https://youtu.be/G_LK0JvdPmc

Detailed Project Description

What exactly does it do?

The crop assistant identifies patterns, through a trained artificial intelligence, some of its features are: identify patterns of infestation levels, identify the plants that are being affected by insects, say what type of insect it may be, and the assistant of Crop 4.0 also issued a warning to the producer when the level of infestation starts to pass the adequate and starts to probably cause serious damages to the productivity of the crop.

How does it work?

The assistant works based on satellite images, such as the launch of the landsat 9, through API (Application Programming Interface), our system will make a service request in the NASA system where the photo data of the region that we are working is stored , through this API then our system gets the necessary images for our AI (Artificial intelligence) which will be trained to identify patterns of infestation for certain cultures.

What benefits does it have?

We can say that an insect becomes a pest when it increases in population, to the point of causing disturbances in crop development, in addition to financial losses. Pest control is something that costs a lot, to get an idea about it only for the control of pests such as Spodoptera frugiperda in corn, the total amount spent in Brazil in a single harvest exceeds R$ 758 million. If we consider that there are two harvests, the cost reaches R$1.2 billion/year. With the 4.0 assistant, these costs would be reduced, as it would help the product to see the area that has the highest level of pest insect infestation and the time to intervene, based on NASA spectral data analysis, the assistant will be able to differentiate between the color of an area of plants with insect infestations, and with the thermal data, you can quantify, based on a specific algorithm for each culture, the time to intervene. Starting control at the right time is essential, as it is not because there are some caterpillars in the crop that you should enter with insecticide, some insects even help in the balance of the agroecosystem. The important thing is to start control before the impacts caused by the pest have reached the level of economic damage.

In addition to identifying the location of the crop that has the problem, the assistant would help identify possible types of pests, this would be done by evaluating the type of bites and the pattern that the insect attacked the plant. Insects have varied eating habits, and also have different mouthparts. So it would be possible to make the difference between a caterpillar, which has a chewing-type mouthparts and a bug that has a sucker-type mouthparts, still talking about caterpillars, have caterpillars that are more generalist and feed on all parts of the leaf, however, there are also those that feed only on the leaf blade and preserve the veins.

What do you hope to achieve?

We hope to be able to reduce losses in agricultural productivity, and thereby help the environment, as the assistant will help to have a more precise control and will not only avoid spending by the producer, but will help to reduce spending on insecticides, and reducing these spending we would also be helping the environment and the final consumer.

What do you hope to achieve? What tools, coding languages, hardware, or software did you use to develop your project?

Python is the programming language that will be used together with the Tensor Flow library.

Space Agency Data

Through API (Application Programming Interface), our system would make a service request in the NASA system where the photo data and thermal data of the region we are serving are stored, through this API then our system gets the necessary images for the our AI already trained with the patterns of infestation and bites of pest insects, detecting the time to intervene and even the invader so that the producer can take the necessary measures.

we use the material available in a videos, Seeing the Unseeable - Viewing Bugs from Space | NASA Space Apps Challenge

Available in: https://youtu.be/yg26sWAm6P4

Hackathon Journey

What did you learn?

I learned a lot, it's interesting to see different points of view.

What inspired your team to choose this challenge?

What inspired the team to choose the SEEING THE UNSEEABLE - VIEWING BUGS FROM SPACE challenge was the affinity with the theme and also the opportunity to bring this to solve solutions for agriculture. Aiming at our region, Mato Grosso, Brazil's main agricultural producer, the problem with insect pests is something highly common and leads to losses, these losses in the crop will often reflect in the final link of the chain, the consumers, often resulting in an increase prices. So, thinking of helping not only the producer but also benefiting the final consumers, we developed the 4.0 crop assistant, an idea that will help the producer in decision making.

What was your approach to developing this project?

The costs with pesticides reach 40% of the cost of production (IMEA), we have a google forms validating the problem in which 50% of producers reported: problems with high costs, low accuracy in identifying symptoms. We watched some nasa space apps challenge videos, and based on them, we were able to follow a line of reasoning that the rest of the group shared. Assistant 4.0 was elaborated based on the experiences that the group members had, Carol has worked with insects and has a good base on this, Ighor and Gabriel took care of the technical part of the project.

How did your team resolve setbacks and challenges?

We solve setbacks and challenges as a group, always based on conversation. The biggest enemy was time, but we tried to optimize by dividing the tasks.