Increase the efficiency of the device to sort apples by recognizing their characteristics

Tsvirkun, Liudmyla and Цвіркун, Людмила Олександрівна and Tsvirkun, Sergij and Цвіркун, Сергій Леонідович (2019) Increase the efficiency of the device to sort apples by recognizing their characteristics. Праці Таврійського державного агротехнологічного університету, 1. pp. 125-131. ISSN 2078-0877

[img]
Preview
Text
Tsvirkun_article_10_05_2019.pdf

Download (479kB) | Preview

Abstract

The article describes the increase in the efficiency of the device for sorting apples by recognizing their characteristics. It is proposed to recognize the contours of objects when they appear in the frame, and further their video surveillance using unique special points of the image. When processing the upper fragment of the fα (x, y, t) frame, the contours of apples and their visual characteristics are determined: size (d), weight (m), color (g), search for special points on the entire image and calculation of their descriptors. The unique descriptors of those special points belonging to the selected objects are remembered and, according to their location in the lower fragment of the frame fβ (x, y, t), the location of the corresponding apples to be selected is evaluated. When forming an automated control over the process of sorting apples, it is advisable to select several samples with better characteristics if their location on the conveyor line does not allow all objects to be taken from the flow. The process of sorting apples is advisable to carry out on the basis of contactless measurement of such characteristics as size, weight, color. It was determined that when managing the process of sorting objects on a conveyor line in the food industry, it is advisable to monitor the video signal on the basis of a paired analysis of sequential frames. At the same time, the trajectory of moving objects in the stream should be tracked based on the calculation of the similarity function between the reference image on the previous one and one of the many fragments lying in the search area on the next frame.

Item Type: Article
Uncontrolled Keywords: sorting device, object recognition, SIFT method, FAST method, apple
Subjects:
Divisions:
Depositing User: Адміністратор репозиторію
Date Deposited: 15 Jul 2019 06:16
Last Modified: 15 Jul 2019 06:16
URI: http://elibrary.donnuet.edu.ua/id/eprint/1467

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year

View more statistics