Autoria: BARBEDO, J. G. A.

Resumo: Abstract-Counting objects is an important activity in the daily routine of many areas of industry. This is particularly true in agriculture, in which objects like cells, microorganisms, seeds and other structures have to be quantified as a source of relevant information. This paper proposes a framework that aggregates three different algorithms into a single tool able to tackle a wide variety of counting problems that exist in the agriculture industry. The factor that brings all those algorithms together is the input by the user of some templates for the objects, which allows the resulting method to select the best option for those particular conditions. As a desirable side effect, problems related to resolution and scale dependencies that plagued those previous algorithms are mostly solved by this new approach.

Ano de publicação: 2012

Tipo de publicação: Artigo em anais e proceedings

Unidade: Embrapa Agricultura Digital

Palavras-chave: Aplicação agrícola, Contagem de objetos, Imagens digitais