CNN UNTUK DETEKSI BOLA MULTI POLA STUDI KASUS : LIGA HUMANOID ROBOCUP

Maulana Aziz Assuja, Achmad Fauzi

Abstract

One of the most popular applications of computer vision in the field of robotics is in the object detection section. The diversity of shapes, patterns, colors, light conditions and occlusion are the main challenges that must be solved. In the case of the humanoid league held by RoboCup, each robot that is required to only use a camera to recognize the surrounding environment. One of the most important objects is, where the rules of the ball are only limited by the 50% rule on a white surface, creating the possibility of a variety of pattern and color combinations that must be handled by the robot vision system. This research was conducted to test the application of the Convolutional Neural Network (CNN) algorithm, especially on the MobileNet-V1 architecture. CNN's implementation is carried out using tensorflow with a dataset of 50,000 labeled instants taken using a PSEye camera with a resolution of 640x480 pixels. Based on testing, CNN is able to handle well variations in object and background conditions with the best accuracy reaching 96%.

Keywords

CNN; Ball Detection; Object Detection; MobileNet; TensorFlow;

Full Text:

PDF

References

[1] Susanto, E. Rudiawan, R. Analia, P. D. Sutopo, and H. Soebakti, “The deep learning development for real-time ball and goal detection of barelang-FC,” in Proceedings IES-ETA 2017 - International Electronics Symposium on Engineering Technology and Applications, 2017, vol. 2017-Decem, pp. 146–151. doi: 10.1109/ELECSYM.2017.8240393.

[2] M. Teimouri, M. H. Delavaran, and M. Rezaei, “A Real-Time Ball Detection Approach Using Convolutional Neural Networks,” in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2019, vol. 11531 LNAI, pp. 323–336. doi: 10.1007/978-3-030-35699-6_25.

[3] D. Setiawan, I. S. Rosandi, M. Pajar, K. Putra, S. Darmawan, and R. Teknokrat, Deteksi Bola Multipola Pada Robot Krakatau FC. 2017.

[4] N. Khamdi, M. Susantok, and P. Leopard, “PENDETEKSIAN OBJEK BOLA DENGAN METODE COLOR FILTERING HSV PADA ROBOT SOCCER HUMANOID,” vol. 6, no. 2, 2017, doi: 10.20449/jnte.v6i2.398.

[5] M. P. K. Putra, “Deteksi Bola Multipola Memanfaatkan Ekstraksi Fitur Local Binary Pattern dengan Algoritma Learning Adaboost,” vol. 1, no. 1, pp. 118–122, 2021.

[6] F. N. Iswahyudi, B. Alldino, and A. Sumbodo, “Pendeteksian Bola untuk Robot Sepak Bola Humanoid Berbasis Pengenalan Pola,” IJEIS, vol. 7, no. 1, pp. 105–116, 2017.

[7] K. P. Danikusumo, “IMPLEMENTASI DEEP LEARNING MENGGUNAKAN CONVOLUTIONAL NEURAL NETWORK UNTUK KLASIFIKASI CITRA CANDI BERBASIS GPU.” 2017.

[8] S. R. Dewi, “DEEP LEARNING OBJECT DETECTION PADA VIDEO.” 2018.

[9] F. Chollet and others, “Keras.” 2015.

[10] A. Paszke et al., “PyTorch: An Imperative Style, High-Performance Deep Learning Library.” 2019.

[11] Martín~Abadi et al., “TensorFlow: Large-Scale Machine Learning on Heterogeneous Systems.” 2015. [Online]. Available: https://www.tensorflow.org/

[12] S. Albawi, T. A. M. Mohammed, and S. Alzawi, “Layers of a Convolutional Neural Network,” Ieee, p. 16, 2017.

[13] S. F. Sindy, “PENDETEKSIAN OBJEK MANUSIA SECARA REAL TIME DENGAN METODE MOBILENET-SSD MENGGUNAKAN MOVIDIUS NEURAL COMPUTE STICK PADA RASPBERRY PI.” 2019.

Refbacks

  • There are currently no refbacks.