SEPTIAWAN, BAYU (2025) IMPLEMENTASI ALGORITMA DETEKSI OBJEK DAN OPTICAL CHARACTER RECOGNITION (OCR) PADA ROBOT LENGAN 3 DEGREE OF FREEDOM (DOF). S1 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
01 COVER.pdf Download (2MB) | Preview |
|
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (BAB I)
02 BAB 1.pdf Restricted to Registered users only Download (122kB) |
|
|
Text (BAB II)
03 BAB 2.pdf Restricted to Registered users only Download (330kB) |
|
|
Text (BAB III)
04 BAB 3.pdf Restricted to Registered users only Download (293kB) |
|
|
Text (BAB IV)
05 BAB 4.pdf Restricted to Registered users only Download (390kB) |
|
|
Text (BAB V)
06 BAB 5.pdf Restricted to Registered users only Download (181kB) |
|
|
Text (DAFTAR PUSTAKA)
07 DAFTAR PUSTAKA.pdf Restricted to Registered users only Download (128kB) |
|
|
Text (LAMPIRAN)
08 LAMPIRAN.pdf Restricted to Registered users only Download (329kB) |
Abstract
This study aims to develop an automation system based on Optical Character Recognition (OCR) technology integrated with a 3 Degree of Freedom (DOF) Robotic Arm. The system uses a Raspberry Pi 3 B+ as the main processing unit and a camera as the visual sensor to capture images of labeled objects. Character recognition is performed using the Tesseract OCR algorithm supported by image pre-processing through OpenCV. The recognized text results are transmitted to an Arduino Uno via serial communication to control the Robotic Arm's servo motors using the inverse kinematics method. System testing was conducted on four objects with different text labels (Amazon, Tokopedia, Bimoli, and Cocacola) to evaluate character recognition accuracy, response time, and the influence of distance, camera angle, and lighting conditions. The output validation used the difflib string similarity method to match recognized text with predefined keywords, which effectively enhanced system reliability against minor recognition errors. With an average recognition and execution time of under 4 seconds, the system is responsive enough for simple real-time applications. The two-stage OCR approach (OCR 1 and OCR 2) proves to be complementary in enhancing text recognition quality. Overall system integration testing—which includes object detection, text recognition, data transmission to the Arduino, and robotic motion execution shows an average success rate of 75%, where success is defined by the robot correctly performing actions based on the recognized text. The success rate varies by object: Amazon (100%), Tokopedia (50%), Bimoli (30%), and Cocacola (0%). The results showed the highest accuracy of 80% on objects with simple fonts and bright lighting. In contrast, accuracy significantly decreased on decorative fonts, tilted angles, non-optimal distances, and low-light conditions. This research demonstrates that the integration of OCR, image processing, and inverse kinematics-based robotic motion control can realize an adaptive and efficient textbased automatic object sorting system, contributing meaningfully to the development of computer vision-based robotics. Keywords: OCR, Raspberry Pi, Tesseract, 3-DOF Robotic Arm, difflib, inverse kinematics, OpenCV, image processing. Penelitian ini bertujuan untuk mengembangkan sistem otomatisasi berbasis teknologi Optical Character Recognition (OCR) yang terintegrasi dengan robot lengan 3 Degree of Freedom (DOF). Sistem ini menggunakan Raspberry Pi 3 B+ sebagai unit pemrosesan utama dan kamera sebagai sensor visual untuk menangkap citra objek berlabel teks. Proses pengenalan karakter dilakukan melalui algoritma Tesseract OCR dengan dukungan pra-pemrosesan citra menggunakan OpenCV. Hasil pembacaan teks dikirim ke Arduino Uno melalui komunikasi serial untuk menggerakkan motor servo pada lengan robot menggunakan metode inverse kinematics. Pengujian sistem dilakukan pada empat objek dengan label berbeda (Amazon, Tokopedia, Bimoli, dan Cocacola) untuk mengevaluasi akurasi pembacaan karakter, waktu respon, dan pengaruh variasi jarak, sudut kamera, serta kondisi pencahayaan. Sebaliknya, akurasi menurun drastis pada font dekoratif, sudut kamera miring, jarak tidak ideal, serta pencahayaan rendah. Validasi hasil OCR dilakukan menggunakan metode difflib untuk membandingkan keluaran teks dengan kata kunci target, yang terbukti meningkatkan keandalan sistem dalam menghadapi kesalahan minor. Dengan waktu rata-rata pengenalan dan eksekusi di bawah 4 detik, sistem ini cukup responsif untuk aplikasi real-time sederhana. Pendekatan dua tahap OCR (OCR 1 dan OCR 2) terbukti saling melengkapi dalam meningkatkan kualitas pengenalan teks. Pengujian integrasi sistem secara keseluruhan yang mencakup deteksi objek, pembacaan teks, pengiriman data ke Arduino, dan eksekusi gerakan robot menunjukkan tingkat keberhasilan rata-rata sebesar 75%, dengan kriteria keberhasilan ditentukan dari kesesuaian gerakan robot terhadap hasil pembacaan teks. Adapun hasil per objek menunjukkan variasi, yaitu Amazon (100%), Tokopedia (50%), Bimoli (30%), dan Cocacola (0%). Hasil menunjukkan bahwa sistem memiliki tingkat akurasi tertinggi sebesar 80% pada objek dengan font sederhana dan pencahayaan terang. Penelitian ini menunjukkan bahwa integrasi antara OCR, image processing, dan kendali gerak robot berbasis inverse kinematics dapat mewujudkan sistem pemilah objek otomatis berbasis teks yang adaptif dan efisien, serta memberikan kontribusi nyata dalam pengembangan robotika berbasis visi komputer. Kata kunci: OCR, Raspberry Pi, Tesseract, robot lengan 3 DOF, difflib, inverse kinematics, OpenCV, pengolahan citra.
Actions (login required)
 |
View Item |
