BerryIP embedded: An embedded vision system for strawberry crop

• Published in: Computers and electronics in agriculture Vol. 173; p. 105354
• Main Authors: de Castro, Andreison, Madalozzo, Guilherme Afonso, dos Santos Trentin, Nicolas, Castoldi da Costa, Rosiane, Calvete, Eunice Oliveira, Schardong Spalding, Luiz Eduardo, Rieder, Rafael
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.06.2020; Elsevier BV
• Subjects: Coefficient of variation; Computer vision; Data analysis; Embedded system; Embedded systems; Equipment costs; Graphical user interface; Image manipulation; Image processing; Leaf area; Meteorological data; Software; Strawberries; Strawberry crop; Vision systems; Weather sensors; 99-00; Computer vision; 00-01; Embedded system; Leaf area; Weather sensors; Strawberry crop
• ISSN: 0168-1699; 1872-7107
• DOI: 10.1016/j.compag.2020.105354

•Creation of a new embedded vision system for the strawberry crop: BerryIP Embedded.•Software to view the crop and operate remotely the system, and organize stats.•Show a method to determine the strawberry leaf area using image processing techniques.•Results show our cost-effective system is useful to high-throughput phenotyping.•BerryIP Embedded allows integration with weather sensors and third party systems. The aim of this work is to present an embedded vision system for the strawberry crop named “berryIP Embedded”. We developed a complete solution, considering hardware with sensors, camera and wi-fi in an embedded platform, sending information to software to collect weather data and to determine the leaf area by image manipulation techniques. This software also presents crop weather and image results in a graphical user interface, allowing the system operation for distance and generating statistical data for crop analysis. We used a indoor greenhouse at the University of Passo Fundo to validate the equipment. Results suggested our cost-effective system that could be used in practice by researchers, allowing an effective monitoring of the crop. Data collections were performed during the 21 days, and the data obtained were statistically analyzed. A comparison was executed between the manual method of estimating leaf area of Albion culture, through prediction equations, and the proposed method of image processing, showing that data measured by the platform does not exceed 10% variation. Pearson’s coefficient showed a strong correlation (ρ=0.96) between leaf area and accumulated temperature during the period.