Skip to main content

ID Card Auto-Capture

The SDK automates the entire Thai National ID card photo workflow: it detects the card boundary in the live camera feed, waits for a sharp and stable frame, perspective-corrects the crop, and submits it to the Thai National ID Card OCR API — 1.25 IC per front page, 0.75 IC per back page. No manual framing, no blurry retakes.

Flutter​

import 'package:iapp_ekyc_sdk/iapp_ekyc_sdk.dart';

final client = IappEkycClient(apiKey: 'YOUR_API_KEY');

// Front side
final front = await DocumentCaptureView.start(
  context,
  client: client,
  documentType: DocumentType.thaiIdFront,
  locale: EkycLocale.th,
);

// Back side (laser code)
final back = await DocumentCaptureView.start(
  context,
  client: client,
  documentType: DocumentType.thaiIdBack,
);

Web​

import { IappEkyc } from '@iapp-technology/ekyc-sdk';

const ekyc = new IappEkyc({ apiKey: 'YOUR_API_KEY' });

const front = await ekyc.captureDocument({
  mount: document.getElementById('ekyc-mount'),
  documentType: 'thaiIdFront',
  locale: 'th',
});

const back = await ekyc.captureDocument({
  mount: document.getElementById('ekyc-mount'),
  documentType: 'thaiIdBack',
});

The result exposes typed accessors plus the raw OCR response (result.raw) — see the Thai National ID Card OCR field reference.

iOS / Android / React Native​

The same flow ships in the native wrappers — one engine, identical capture quality. Install and camera-permission setup: iOS, Android & React Native.

// iOS (Swift; Objective-C also supported)
let config = IappEkycConfig(apiKey: "YOUR_API_KEY", flow: .documentCapture)
config.documentType = .thaiIdFront
IappEkycSdk.present(from: self, config: config) { result in /* result.document?.rawJSON */ }
// Android (Kotlin; Java also supported)
ekyc.launch(IappEkycRequest.DocumentCapture(config, EkycDocumentType.THAI_ID_FRONT))
// React Native
<IappEkycFlow flow="documentCapture" documentType="thaiIdFront" apiKey="YOUR_API_KEY"
  onResult={handleResult} onError={handleError} onCancel={close} />

How auto-capture works​

  • Boundary detection — adaptive Canny edge detection and contour analysis (OpenCV) find a convex quadrilateral matching the ID-1 card aspect ratio (1.586, i.e. 85.60 × 53.98 mm, ±0.25 tolerance) inside the on-screen guide.
  • Stability — capture only triggers after the card has been steady across a sliding window of recent frames (corner drift under 2% of the frame diagonal).
  • Sharpness — Laplacian-variance scoring rejects blurry frames before any credit is spent.
  • Perspective correction — the detected corners are warped to a flat ~300 DPI card image (1011 × 637 px) before upload, which measurably improves OCR accuracy.
  • Manual fallback — a manual shutter button appears after 10 seconds if auto-capture has not fired.

Learn more​