Derin Öğrenme ile Nesne Tanıyan Robot Uygulaması
Yıl 2021,
Sayı: 31, 127 - 133, 31.12.2021
Uğur Talaş
,
Uğur Yüzgeç
,
Burakhan Çubukçu
Öz
Günümüzde birçok farklı alanda insanlara fayda sağlamak için teknolojik cihazlar ve robotlar kullanılmaktadır. Özellikle askeri alanda insan hayatının riske girebileceği ortamlarda robotlar yardımıyla, hayati riskler minimize edilmek istenilmektedir. Askeri operasyonlarda bir binaya keşif amaçlı bir insanın girmesi oldukça riskli bir durumdur. Bu çalışmada bu tür riskli durumlarda insanın keşif yapması yerine uzaktan kontrol edilebilen, gördüğü nesneleri tanıyabilen ve tanıdığı nesneleri kontrol ekranında gösteren bir robot tasarlanmıştır. Bu çalışmada geliştirilen robot nesne tanımak için Google tarafından geliştirilen TensorFlow derin öğrenme kütüphanesini kullanmaktadır. Python diliyle geliştirilen yazılım robot üzerinde bulunan Raspberry Pi3/B mini bilgisayarı üzerinde çalıştırılmıştır. Robot hareketi için DC motorlardan faydalanılmıştır. Raspberry Pi3/B mini bilgisayarı üzerindeki GPIO pinleri ile motor sürücü devresine sinyal gönderilerek robotun hareketlerinin kontrol edilebilmesi sağlanılmıştır. Yapılan prototipin testlerinde nesneleri çoğunlukla başarılı şekilde tanınabildiği ve uygun ışık ortamında başarı oranının arttığı gözlemlenmiştir.
Kaynakça
- Abadi, M., Agarwal, A., Barham, P., Brevdo, E., Chen, Z., Citro, C., … Zheng, X. (2016). TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems.
- Abadi, M., Barham, P., Chen, J., Chen, Z., Davis, A., Dean, J., … Brain, G. (2016). TensorFlow: A System for Large-Scale Machine Learning.
- Ataner, E., Özdeş, B., Öztürk, G., Yasin, T., Çelik, C., Durdu, A., … Öz, Y. (2020). Deep Learning Methods in Unmanned Underwater Vehicles. European Journal of Science and Technology Special Issue, 345–350. doi:10.31590/ejosat.804599
- Aydın, Z. (2020). Performance Analysis of Machine Learning and Bioinformatics Applications on High Performance Computing Systems. Academic Platform Journal of Engineering and Science, 8(1), 1–14. doi:10.21541/apjes.547016
- Celik, Y. ve Güneş, M. (2018). Designing an Object Tracker Self-Balancing Robot. Academic Platform Journal of Engineering and Science, 6(2), 124–133. doi:10.21541/apjes.414715
- Dey, A. (2016). Machine Learning Algorithms: A Review. International Journal of Computer Science and Information Technologies, 7(3), 1174–1179.
- E. R. Davies. (2005). Machine Vision: Theory, Algorithms, Practicalities.
- Girshick, R., Donahue, J., Darrell, T. ve Malik, J. (y.y.). Rich feature hierarchies for accurate object detection and semantic segmentation.
- KOCAMAZ, A. F. (2012). Makine Öğrenmesi Tabanlı Bir Uzman Sistem Tasarımı.
- KÖSE, U. (2017). YAPAY ZEKÂ TABANLI OPTİMİZASYON ALGORİTMALARI GELİŞTİRİLMESİ, 4, 9–15.
- Krishnan˚, S. K., Fox˚, R. F. ve Goldberg, K. (2017). DDCO: Discovery of Deep Continuous Options for Robot Learning from Demonstrations.
- McCarthy, J., Minsky, M. L., Rochester, N. ve Shannon, C. E. (2006). A proposal for the Dartmouth summer research project on artificial intelligence. AI Magazine, 27(4), 12–14.
- Object Detection using Fast R CNN | Azure AI Gallery. (2021). 15 Mart 2021 tarihinde https://gallery.azure.ai/Tutorial/Object-Detection-using-Fast-R-CNN-1 adresinden erişildi.
- Parvat, A., Chavan, J., Kadam, S., Dev, S. ve Pathak, V. (2017). A survey of deep-learning frameworks. Proceedings of the International Conference on Inventive Systems and Control, ICISC 2017 içinde . Institute of Electrical and Electronics Engineers Inc. doi:10.1109/ICISC.2017.8068684
- Pena, D., Forembski, A., Xu, X. ve Moloney, D. (y.y.). Benchmarking of CNNs for Low-Cost, Low-Power Robotics Applications.
- Phon-Amnuaisuk, S., Murata, K. T., Pavarangkoon, P., Yamamoto, K. ve Mizuhara, T. (2018). Exploring the Applications of Faster R-CNN and Single-Shot Multi-box Detection in a Smart Nursery Domain.
- Puthussery, A. R., Haradi, K. P., Erol, B. A., Benavidez, P., Rad, P. ve Jamshidi, M. (2017). A deep vision landmark framework for robot navigation. 2017 12th System of Systems Engineering Conference, SoSE 2017 içinde . Institute of Electrical and Electronics Engineers Inc. doi:10.1109/SYSOSE.2017.7994976
- Sağlam, A., Taş, M., Baykan, N. A., Üniversitesi, K. T., Ve Doğa, M., Fakültesi, B., … Konya, T. (2020). Geri Dönüştürülebilir Atıkların Materyallerine Göre Sınıflandırılması için Raspberry Pi Tabanlı Donanım Geliştirilmesi. European Journal of Science and Technology Special Issue, 30–38. doi:10.31590/ejosat
- Shaoqing Ren, Kaiming He, Ross Girshick, and J. S. (2017). IEEE Xplore Full-Text PDF: Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks içinde . https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=7485869 adresinden erişildi.
- Shatnawi, A., Al-Bdour, G., Al-Qurran, R. ve Al-Ayyoub, M. (2018). A comparative study of open source deep learning frameworks. 2018 9th International Conference on Information and Communication Systems, ICICS 2018 içinde (C. 2018-January, ss. 72–77). Institute of Electrical and Electronics Engineers Inc. doi:10.1109/IACS.2018.8355444
- Tensorflow. (2019). GitHub - tensorflow/models: Models and examples built with TensorFlow. 15 Mart 2021 tarihinde https://github.com/tensorflow/models adresinden erişildi.
- TensorFlow. (2021). 15 Mart 2021 tarihinde https://www.tensorflow.org/ adresinden erişildi.
- Vassili Kovalev, Alexander Kalinovsky ve Sergey Kovalev. (2016). (PDF) Deep Learning with Theano, Torch, Caffe, TensorFlow, and Deeplearning4J: Which One Is the Best in Speed and Accuracy? (ss. 99–103).
- Yang, G., Yang, J., Sheng, W., Erivaldo Fernandes Junior, F. ve Li, S. (2018). Convolutional Neural Network-Based Embarrassing Situation Detection under Camera for Social Robot in Smart Homes. doi:10.3390/s18050530
- Yapıcı, M. M. ve Topaloğlu, N. (2021). Performance comparison of deep learning frameworks. Computers and Informatics (C. 1).
Object Recognizing Robot Application with Deep Learning
Yıl 2021,
Sayı: 31, 127 - 133, 31.12.2021
Uğur Talaş
,
Uğur Yüzgeç
,
Burakhan Çubukçu
Öz
Today, technological devices and robots are used to benefit in many different areas. In environments where people of other military field may be at risk, vital risks are desired to be minimized for robots. It is very risky for a person to enter a building for reconnaissance purposes during military operations. It is the place where this kind of risky person learns a robot that can be remotely controlled, recognize the text he sees, and display the text control text he knows, instead of making exploration. In this structure, the robot uses the TensorFlow deep learning library offered by Google to recognize objects. It was run on the Raspberry Pi3/B minicomputer on the software with the language of Python. DC motors are used for robot movement. In Raspberry Pi3/B minicomputer, the robot's movements can be controlled by sending a signal to the motor driver circuit with GPIO pins. In the tests of the prototype, it has been observed that the guarantee of success in the distribution area has increased.
Kaynakça
- Abadi, M., Agarwal, A., Barham, P., Brevdo, E., Chen, Z., Citro, C., … Zheng, X. (2016). TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems.
- Abadi, M., Barham, P., Chen, J., Chen, Z., Davis, A., Dean, J., … Brain, G. (2016). TensorFlow: A System for Large-Scale Machine Learning.
- Ataner, E., Özdeş, B., Öztürk, G., Yasin, T., Çelik, C., Durdu, A., … Öz, Y. (2020). Deep Learning Methods in Unmanned Underwater Vehicles. European Journal of Science and Technology Special Issue, 345–350. doi:10.31590/ejosat.804599
- Aydın, Z. (2020). Performance Analysis of Machine Learning and Bioinformatics Applications on High Performance Computing Systems. Academic Platform Journal of Engineering and Science, 8(1), 1–14. doi:10.21541/apjes.547016
- Celik, Y. ve Güneş, M. (2018). Designing an Object Tracker Self-Balancing Robot. Academic Platform Journal of Engineering and Science, 6(2), 124–133. doi:10.21541/apjes.414715
- Dey, A. (2016). Machine Learning Algorithms: A Review. International Journal of Computer Science and Information Technologies, 7(3), 1174–1179.
- E. R. Davies. (2005). Machine Vision: Theory, Algorithms, Practicalities.
- Girshick, R., Donahue, J., Darrell, T. ve Malik, J. (y.y.). Rich feature hierarchies for accurate object detection and semantic segmentation.
- KOCAMAZ, A. F. (2012). Makine Öğrenmesi Tabanlı Bir Uzman Sistem Tasarımı.
- KÖSE, U. (2017). YAPAY ZEKÂ TABANLI OPTİMİZASYON ALGORİTMALARI GELİŞTİRİLMESİ, 4, 9–15.
- Krishnan˚, S. K., Fox˚, R. F. ve Goldberg, K. (2017). DDCO: Discovery of Deep Continuous Options for Robot Learning from Demonstrations.
- McCarthy, J., Minsky, M. L., Rochester, N. ve Shannon, C. E. (2006). A proposal for the Dartmouth summer research project on artificial intelligence. AI Magazine, 27(4), 12–14.
- Object Detection using Fast R CNN | Azure AI Gallery. (2021). 15 Mart 2021 tarihinde https://gallery.azure.ai/Tutorial/Object-Detection-using-Fast-R-CNN-1 adresinden erişildi.
- Parvat, A., Chavan, J., Kadam, S., Dev, S. ve Pathak, V. (2017). A survey of deep-learning frameworks. Proceedings of the International Conference on Inventive Systems and Control, ICISC 2017 içinde . Institute of Electrical and Electronics Engineers Inc. doi:10.1109/ICISC.2017.8068684
- Pena, D., Forembski, A., Xu, X. ve Moloney, D. (y.y.). Benchmarking of CNNs for Low-Cost, Low-Power Robotics Applications.
- Phon-Amnuaisuk, S., Murata, K. T., Pavarangkoon, P., Yamamoto, K. ve Mizuhara, T. (2018). Exploring the Applications of Faster R-CNN and Single-Shot Multi-box Detection in a Smart Nursery Domain.
- Puthussery, A. R., Haradi, K. P., Erol, B. A., Benavidez, P., Rad, P. ve Jamshidi, M. (2017). A deep vision landmark framework for robot navigation. 2017 12th System of Systems Engineering Conference, SoSE 2017 içinde . Institute of Electrical and Electronics Engineers Inc. doi:10.1109/SYSOSE.2017.7994976
- Sağlam, A., Taş, M., Baykan, N. A., Üniversitesi, K. T., Ve Doğa, M., Fakültesi, B., … Konya, T. (2020). Geri Dönüştürülebilir Atıkların Materyallerine Göre Sınıflandırılması için Raspberry Pi Tabanlı Donanım Geliştirilmesi. European Journal of Science and Technology Special Issue, 30–38. doi:10.31590/ejosat
- Shaoqing Ren, Kaiming He, Ross Girshick, and J. S. (2017). IEEE Xplore Full-Text PDF: Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks içinde . https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=7485869 adresinden erişildi.
- Shatnawi, A., Al-Bdour, G., Al-Qurran, R. ve Al-Ayyoub, M. (2018). A comparative study of open source deep learning frameworks. 2018 9th International Conference on Information and Communication Systems, ICICS 2018 içinde (C. 2018-January, ss. 72–77). Institute of Electrical and Electronics Engineers Inc. doi:10.1109/IACS.2018.8355444
- Tensorflow. (2019). GitHub - tensorflow/models: Models and examples built with TensorFlow. 15 Mart 2021 tarihinde https://github.com/tensorflow/models adresinden erişildi.
- TensorFlow. (2021). 15 Mart 2021 tarihinde https://www.tensorflow.org/ adresinden erişildi.
- Vassili Kovalev, Alexander Kalinovsky ve Sergey Kovalev. (2016). (PDF) Deep Learning with Theano, Torch, Caffe, TensorFlow, and Deeplearning4J: Which One Is the Best in Speed and Accuracy? (ss. 99–103).
- Yang, G., Yang, J., Sheng, W., Erivaldo Fernandes Junior, F. ve Li, S. (2018). Convolutional Neural Network-Based Embarrassing Situation Detection under Camera for Social Robot in Smart Homes. doi:10.3390/s18050530
- Yapıcı, M. M. ve Topaloğlu, N. (2021). Performance comparison of deep learning frameworks. Computers and Informatics (C. 1).