RGB Images Driven Recognition of Grapevine Varieties

článek ve sborníku ve WoS nebo Scopus

en

We present a grapevine variety recognition system based on a densely connected convolutional network. The proposed solution is aimed as a data processing part of an affordable sensor for selective harvesters. The system classifies size normalized RGB images according to varieties of grapes captured in the images. We train and evaluate the system on in-field images of ripe grapes captured without any artificial lighting, in a direction of sunshine likewise in the opposite direction. A dataset created for this purpose consists of 7200 images classified into 8 categories. The system distinguishes among seven grapevine varieties and background, where four and three varieties have red and green grapes, respectively. Its average per-class classification accuracy is at 98.10 % and 97.47 % for red and green grapes, respectively. The system also well differentiates grapes from background. Its overall average per-class accuracy is over 98 %. The evaluation results show that conventional cameras in combination with the proposed system allow construction of affordable automatic selective harvesters.

We present a grapevine variety recognition system based on a densely connected convolutional network. The proposed solution is aimed as a data processing part of an affordable sensor for selective harvesters. The system classifies size normalized RGB images according to varieties of grapes captured in the images. We train and evaluate the system on in-field images of ripe grapes captured without any artificial lighting, in a direction of sunshine likewise in the opposite direction. A dataset created for this purpose consists of 7200 images classified into 8 categories. The system distinguishes among seven grapevine varieties and background, where four and three varieties have red and green grapes, respectively. Its average per-class classification accuracy is at 98.10 % and 97.47 % for red and green grapes, respectively. The system also well differentiates grapes from background. Its overall average per-class accuracy is over 98 %. The evaluation results show that conventional cameras in combination with the proposed system allow construction of affordable automatic selective harvesters.

Recognition of grapevine varieties; Densely connected convolutional network; In-field images; Agriculture mechanization

29.09.2020

Springer International Publishing

Cham

978-3-030-57802-2

2194-5365

15th International Conference on Soft Computing Models in Industrial and Environmental Applications (SOCO 2020)

1268

216–225

10