Powered by OpenAIRE graph
Found an issue? Give us feedback
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/ LAReferencia - Red F...arrow_drop_down
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
versions View all 1 versions
addClaim

Inteligência artificial utilizada na identificação de espécies e prognóstico visual de mudas florestais

Authors: Leme, Mateus de Campos;

Inteligência artificial utilizada na identificação de espécies e prognóstico visual de mudas florestais

Abstract

A aplicação de técnicas de inteligência artificial é ampla e envolve ciências políticas, econômicas, geopolíticas, sociais, agrárias, ambientais, legais e outras. Nas agrárias, elas são utilizadas e experimentadas no reconhecimento de pragas, classificação de culturas, contagem de indivíduos arbóreos visando a otimização e redução de custos. Dentre estas técnicas, o deep learning, ou aprendizado profundo de máquinas, é uma opção interessante. Também conhecida como redes neurais profundas convolucionais, tem por base modelos matemáticos capazes de realizar o processo de classificação de imagens. No setor florestal, essa metodologia pode ser aplicada visando a melhoria de produtividade e auxiliando na conservação. O presente estudo teve como objetivo aplicar deep learning para classificar imagens de mudas florestais no processo de identificação de mudas e prognóstico visual. O estudo envolveu a criação de dois datasets: um com 610 imagens, sendo 400 para treino, 200 para teste e 10 para validação, balanceado entre as classes: pau d’alho e ipê-verde; e o outro com imagens de cabreúva, pau d’alho, brinco-de-índio, jatobá e ipê-verde, composto por 1298 imagens sendo 1022 para treino, 256 para teste e 20 para validação. As bibliotecas importadas para realização do processamento e classificação foram Keras, com o framework do Tensorflow, e Matplotlib, todas usando a linguagem de programação Python. A avaliação da performance preditiva do modelo utilizou o 10-fold cross validation. Na classificação entre as mudas de pau d’alho e ipê-verde, a acurácia média foi de 96,84%. Para o teste de prognóstico visual, as classes foram divididas em mudas com anomalia e mudas padrão, e a acurácia média foi de 49,18%. Assim, para diferenciar as duas espécies, o modelo pode ser incorporado em uma API ou Software para aplicação em viveiros florestais. Para utilização com fins de realização do prognóstico visual, serão necessárias adaptações.

The application of artificial intelligence techniques is broad and involves political, economic, geopolitical, social, agrarian, environmental, legal, and other sciences. In the agrarian sciences, they are used and experimented with within the recognition of pests, classification of crops, counting of tree individuals aiming at optimization, and cost reduction. Among these techniques, deep learning or deep machine learning is an interesting option. Also known as convolutional deep neural networks, it is based on mathematical models capable of performing the image classification process. In the forestry sector, this methodology can be applied to improve productivity and aid in conservation. The present study aimed to apply deep learning to classify images of forest seedlings in the process of seedling identification and visual prognosis. The study involved the creation of two datasets: one with 610 images, 400 for training, 200 for testing and 10 for validation, balanced between the classes: pau d'alho and ipê-verde and the other with images of cabreúva, pau d'alho, brinco- de-índio, jatobá, and ipê-verde, composed of 1298 images, 1022 for training, 256 for testing and 20 for validation. The libraries imported for processing and classification were Keras, with the Tensorflow framework, and matplotlib, all using the python programming language. The evaluation of the predictive performance of the model used 10-fold Cross-Validation. In the classification between pau d'alho and ipê verde seedlings, the average accuracy was 96.84%. For the visual prediction test, the classes were divided into seedlings with anomaly and standard seedlings, and the average accuracy was 49.18%. Thus, to differentiate the two species, the model can be incorporated into an API or Software for application in forest nurseries. For use for visual prediction purposes, adaptations will be required.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Pós-graduação em Agronomia (Energia na Agricultura) - FCA

001

Country
Brazil
Keywords

nurseries, convolutional neural networks, inteligência artificial, viveiros florestais, redes neurais convolucionais, artificial intelligence

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    0
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Average
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Average
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Average
Powered by OpenAIRE graph
Found an issue? Give us feedback
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
0
Average
Average
Average
Green
Related to Research communities