Надіслати статтю
вул. Інститутська 11, м. Хмельницький, 29016
visnyk.khnu@khmnu.edu.ua
(0382) 67-51-08

МОНІТОРИНГ РОЗВИТКУ КУРЯЧИХ ЯЄЦЬ НА БАЗІ НЕЙРОННОЇ МЕРЕЖІ РОЗПІЗНАВАННЯ СТАНУ ЕМБРІОНІВ

MONITORING THE CHICKEN EGGS DEVELOPMENT ON THE BASIS OF THE NEURAL NETWORK FOR RECOGNITION OF THE EMBRYOS STATE

Сторінки: 95-101. Номер: №6, 2021 (303) 
Автори:
УТКІНА Т. Ю.
Черкаський державний технологічний університет
ORCID ID: 0000-0002-6614-4133
e-mail: t.utkina@chdtu.edu.ua
РЯБЦЕВ В. Г.
ООО “ДП СВ “Альтера”, м. Черкаси
ORCID ID: 0000-0002-0592-2413
TETYANA UTKINA
Cherkasy State Technological University
VLADIMIR RYABTSEV
“LLC DP SV “Altera”, Cherkasy
DOI: https://www.doi.org/10.31891/2307-5732-2021-303-6-95-101
Рецензія/Peer review : 09.11.2021 р.
Надрукована/Printed : 30.12.2021 р.

Анотація мовою оригіналу

У роботі наведено результати досліджень штучної нейронної мережі для моніторингу розвитку курячих яєць з розпізнаванням стану ембріонів, що дозволяє уникнути недоцільного використання інкубатора й відібрати життєздатні ембріони для виведення пташенят. Застосування засобів штучного інтелекту розширює можливості виявлення дефектних яєць, підвищує точність та швидкість візуального контролю під час виконання автоматизованого овоскопування.
Ключові слова: інкубація, штучна нейронна мережа, овоскопування, технічний зір, ембріон.

Розширена анотація англійською мовою

Currently, a lot of human resources are required to get rid on poultry farms from viruses and determining the fertility of chicken embryos with three states: live, dead and weak, and then take measures to increase hatching. The article discusses an artificial neural network (ANN) structure to automate the monitoring of the chicken eggs development with identifying viable embryos. A technique for separating the chicken eggs area to sectors to automate the ovoscoping process has been developed. The structural scheme of the artificial neural network contains a set of synapses, each of which is characterized by its weight. Synapses input signals are multiplied by weight and folded by adder. When adding, a threshold is also taken into account, which has a negative value. To generate a signal at the network output, the activation function of a single jump is applied. A VHDL model of neural network is developed, consisting of an adder of the input signal elements and subtractor vector, discharge grid output signal of which is the output signal of the network. The model of an adder of the input signals vector is designed in VHDL. The modeling of recognition of various states of chicken embryos is performed. When perform the ovoscoping, the light sector of the egg area corresponds to a signal equal to 0, and the dark sector to signal 1. Total, the state of 6 sectors of area of the poultry egg in the model is analyzed. During the simulation, when a live embryo is fixed, the signal at the output of the network is 1, and when a dead egg is detected, this signal is 0. Verification of the VHDL model against test sets is confirmed the diagnostic properties of the ANN model. By including artificial intelligence in the FH vision system, visual control of the chicken embryos development is automated using an autonomous, easily integrated solution, as opposed to executing solutions that require special programming tools and runtime environments. This vision system also does not require artificial intelligence expertise to set up in poultry farms.
Keywords: incubation, artificial neural network, ovoscoping, technical vision, embryo.

References

  1. Fasenko G. M. Egg Storage and the Embryo / G. M. Fasenko // Poultry Science. – 2007. – Vol. 86. – P. 1020–1024.
  2. Hrechanov O. P. Efektyvni rezhymy osvitlennia u ptashnyku / O. P. Hrechanov // Suchasne ptakhivnytstvo. – 2005. – № 7. – S. 16–18.
  3. Morris T. R. Effect of Age at Starting Biomittent Lighting on Performance of Laying Hens / T. R. Morris, M. Midgley, E. A. Butler // British Poultry Science. – 1990. – Vol. 31, No. 3. – P. 447–455.
  4. Hulet R. M. Symposium: Managing the Embryo for Performance Managing Incubation: Where Are We and Why? / R. M. Hulet // Poultry Science. – 2007. – Vol. 86, Is. 5. – P. 1017–1019.
  5. Surface Imaging Microscopy, an Automated Method for Visualizing Whole Embryo Samples in Three Dimensions at High Resolution / A. Ewald, H. McBride, M. Reddington, S. Fraser, R. Kerschmann // Developmental Dynamics. – 2002. – No. 225. – P. 369–375.
  6. Dechow D. Machine Vision Systems Integration: Deep Learning / D. Dechow. – URL : https://www.qualitymag.com/articles/96663-machine-vision-systems-integration-deep-learning
  7. West P. C. Machine Vision Systems Design: The Basics / P. C. West, D. Dechow. – URL : https://www.qualitymag.com/ articles/96209-machine-vision-systems-design-the-basics
  8. Yu H. Chicken Embryo Fertility Detection Based on PPG and Convolutional Neural Network / H. Yu, G. Wang, Z. Zhao, H. Wang, Z. Wang // Infrared Physics & Technology. – 2019. – Vol. 103. – P. 1350–4495. – DOI : https://doi.org/10.1016/j.infrared.2019.103075
  9. Hajkin S. Nejronnye seti: polnyj kurs / Hajkin ; per. s angl. – 2-e izd., ispr. – M. : OOO “I.D. Vilyams”, 2006. – 1104 s.
  10. Kits FHV7 – Smart Camera / S. Kits. – URL : https://www.appliedautomation.co.uk/X-STK/OMRON/downloads/FHV7%20Starter%20Pack%20Offer_X-STK.pdf
  11. Haefner N. Artificial Intelligence and Innovation Management: A Review, Framework, and Research Agenda / Haefner, J.Wincent, V. Paridac, O. Gassmann // Technological Forecasting and Social Change. – 2021. – Vol. 162. – P. 0040–1625. – DOI : https://doi.org/10.1016/j.techfore.2020.120392
  12. Semenets V. V. Proektuvannia tsyfrovykh system z vykorystanniam movy VHDL : navch. posibnyk / V. V. Semenets, I.Khakhanova, V. I. Khakhanov. – Kharkiv : KhNURE, 2003. – 492 s.
  13. Xilinx Libraries Guide. – URL : http://www.cs.columbia.edu/~sedwards/classes/2005/4840/lib.pdf
  14. Das K. Detecting Fertility of Hatching Eggs using Machine Vision II: Neural Network Classifiers. / K. Das, M. D. Evans // Transactions of the American Society of Agricultural Engineers. – 1992. – Vol. 35, Is. 6. – P. 2035–2041.
  15. Omrons FH Series Vision System Expands Automated Inspections Capabilities with AI-Based Defect Detection Technology. – URL : https://finance.yahoo.com/news/omrons-fh-series-vision-system-134000799.html

Post Author: Горященко Сергій

Translate