Syafei, Farros Hilmi (2026) API Verifikasi Dokumen Pengemudi dan Kendaraan Berbasis OCR untuk Onboarding dan Dispatch Logistik. Project Report. [s.n.]. (Unpublished)
![[thumbnail of 5025201012-Project_Report.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5025201012-Project_Report.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy |
Abstract
Proses onboarding pengemudi dan dispatch armada pada platform logistik pihak ketiga masih bergantung pada pemeriksaan dokumen secara manual sehingga memerlukan waktu lama, rentan terhadap kesalahan, dan menyulitkan proses audit. Kerja praktik ini bertujuan mengembangkan layanan verifikasi dokumen berbasis Optical Character Recognition (OCR) yang mampu mengotomatisasi ekstraksi dan validasi informasi dari dokumen KTP, SIM, STNK, dan KIR untuk mendukung proses operasional perusahaan. Sistem diimplementasikan menggunakan PaddleOCR (PP-OCRv5/v6) sebagai mesin OCR, FastAPI sebagai layanan REST, serta Docker untuk memudahkan proses deployment pada infrastruktur perusahaan. Arsitektur layanan dirancang bersifat stateless sehingga dokumen hanya diproses selama permintaan berlangsung tanpa penyimpanan data identitas. Hasil OCR diproses lebih lanjut melalui deteksi jenis dokumen, ekstraksi field berbasis label dan geometri, validasi pola data, koreksi menggunakan data referensi, serta detektor field berbasis YOLO untuk meningkatkan akurasi pada dokumen KTP dan STNK. Sistem menghasilkan keputusan verifikasi berupa match, no_match, atau needs_review beserta kode alasan yang dapat diintegrasikan dengan sistem operasional. Hasil evaluasi menunjukkan akurasi deteksi jenis dokumen sebesar 95%, pembacaan field yang baik pada dokumen berkualitas tinggi, serta keberhasilan implementasi layanan yang siap diintegrasikan ke dalam proses onboarding dan dispatch di lingkungan perusahaan.
======================================================================================================================================
Driver onboarding and fleet dispatch on third-party logistics platforms still rely on manual document checks, which are time-consuming, prone to error, and difficult to audit. This internship aimed to develop a document verification service based on Optical Character Recognition (OCR) capable of automating the extraction and validation of information from Indonesian identity card (KTP), driving licence (SIM), vehicle registration (STNK), and roadworthiness certificate (KIR) documents in support of the company's operational processes.
The system was implemented using PaddleOCR (PP-OCRv5/v6) as the OCR engine, FastAPI as the REST service, and Docker to simplify deployment on the company's infrastructure. The service architecture was designed to be stateless, so that documents are processed only for the duration of a request and no identity data is stored. The OCR output is processed further through document type detection, label and geometry-based field extraction, data pattern validation, correction against reference data, and a YOLO based field detector to improve accuracy on KTP and STNK documents. The system issues a verification decision of match, no_match, or needs_review together with a machine-readable reason code that can be integrated with operational systems. Evaluation showed a document type detection accuracy of 95%, reliable field reading on good-quality documents, and a successful service implementation that is ready to be integrated into the company's onboarding and dispatch processes.
| Item Type: | Monograph (Project Report) |
|---|---|
| Uncontrolled Keywords: | OCR, verifikasi dokumen, document verification, KYC logistik, PaddleOCR, FastAPI, stateless API |
| Subjects: | T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T58.5 Information technology. IT--Auditing |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis |
| Depositing User: | Farros Hilmi Syafei |
| Date Deposited: | 29 Jul 2026 06:31 |
| Last Modified: | 29 Jul 2026 06:31 |
| URI: | http://repository.its.ac.id/id/eprint/138463 |
Actions (login required)
 |
View Item |