Segmentation of Bean-Plants Using Clustering Algorithms

Abstract

In recent years laser scanning platforms have been proven to be a helpful tool for plants traits analysing in agricultural applications. Three-dimensional high throughput plant scanning platforms provide an opportunity to measure phenotypic traits which can be highly useful to plant breeders. But the measurement of phenotypic traits is still carried out with labor-intensive manual observations. Thanks to the computer vision techniques, these observations can be supported with effective and efficient plant phenotyping solutions. However, since the leaves and branches of some plant types overlap with other plants nearby after a certain period of time, it becomes challenging to obtain the phenotypical properties of a single plant. In this study, it is aimed to separate bean plants from each other by using common clustering algorithms and make them suitable for trait extractions. K-means, Hierarchical and Gaussian mixtures clustering algorithms were applied to segment overlapping beans. The experimental results show that K-means clustering is more robust and faster than the others.

Prediction of Aboveground Biomass of Three Cassava (Manihot esculenta) Genotypes Using a Terrestrial Laser Scanner

“…It is also feasible that advances in machine learning could help to mitigate some of the challenges associated with differentiating between overlapping plants, especially if algorithms are designed specifically for this purpose. Scharr et al have already had success using these methods with leaf segmentation in Arabidopsis and tobacco [30], while Kartal et al have had success segmenting overlapping bean plants using a k-means algorithm [31]. For fields which feature dense planting, breeders may also receive better results by pursing large biomass assessments made through UAVs [32].…”

A Strategic Analytics Using Convolutional Neural Networks for Weed Identification in Sugar Beet Fields

“…Transfer learning is a process that facilitates learning from previous learning, provided there is compatibility between the two learnings (Zhuang et al, 2020). The typical use case of transfer learning is when the data are limited (a thousand photos of weeds) and where you learn from a neural network that has learned to recognize thousands of images (possibly involving different types of weed) classify weed species (Kartal et al, 2020).…”

As Técnicas de Monitoramento por VANT: Uma Revisão dos Resultados Obtidos na Agricultura

“…O uso de VANT na pulverização é considerado uma alternativa para melhorar o trabalho em locais de difícil acesso e a diminuição do risco de intoxicação para o operador é uma grande vantagem, mas a tecnologia ainda apresenta problemas de autonomia, deficiência na eficácia na operação e problemas de usabilidade (KARTAL et al, 2020).…”

“…Novas tecnologias estão incrementando o monitoramento das áreas agrícolas, com a necessidade de aumentar a eficiência dos sistemas produtivos, através de uma análise detalhada de cada metro quadrado da lavoura. Essa precisão de monitoramento deve ser mantida independentemente da escala da produção (KARTAL et al, 2020). Nesse contexto, o veículo aéreo não tripulado (VANT) pode contribuir com aumento da precisão no sistema produtivo na identificação de alvos específicos (HERWITZ et al, 2004).…”