{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "GzFYZYcT9oyb"
   },
   "source": [
    "# RDD matching example (sparse, semi-dense and lightglue)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "97Mbt4a89z3Z"
   },
   "source": [
    "## Initialize RDD"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from RDD.RDD import build\n",
    "from RDD.RDD_helper import RDD_helper\n",
    "from matplotlib import pyplot as plt\n",
    "from time import time\n",
    "\n",
    "RDD_model = build(weights='./weights/RDD-v2.pth')\n",
    "RDD_model.eval()\n",
    "RDD = RDD_helper(RDD_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import cv2\n",
    "\n",
    "def draw_matches(ref_points, dst_points, img0, img1):\n",
    "    \n",
    "    # Prepare keypoints and matches for drawMatches function\n",
    "    keypoints0 = [cv2.KeyPoint(p[0], p[1], 1000) for p in ref_points]\n",
    "    keypoints1 = [cv2.KeyPoint(p[0], p[1], 1000) for p in dst_points]\n",
    "    matches = [cv2.DMatch(i,i,0) for i in range(len(ref_points))]\n",
    "\n",
    "    # Draw inlier matches\n",
    "    img_matches = cv2.drawMatches(img0, keypoints0, img1, keypoints1, matches, None,\n",
    "                                  matchColor=(0, 255, 0), flags=2)\n",
    "\n",
    "    return img_matches\n",
    "\n",
    "\n",
    "def draw_points(points, img):\n",
    "    for p in points:\n",
    "        cv2.circle(img, (int(p[0]), int(p[1])), 2, (0, 255, 0), -1)\n",
    "        \n",
    "    return img\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "b83vE-Dt-cTC"
   },
   "source": [
    "## Matching example - Sparse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Load some example images\n",
    "im0 = cv2.imread('./assets/image0.jpg')\n",
    "im1 = cv2.imread('./assets/image1.jpg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 431
    },
    "id": "8qm_cdIq9-jy",
    "outputId": "ebd99a35-807d-4684-f43b-4f1b0a022c66"
   },
   "outputs": [],
   "source": [
    "start = time()\n",
    "mkpts_0, mkpts_1, conf = RDD.match(im0, im1, resize=1024)\n",
    "print(f\"Found {len(mkpts_0)} matches in {time()-start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "canvas = draw_matches(mkpts_0, mkpts_1, im0, im1)\n",
    "plt.figure(figsize=(12,12))\n",
    "plt.imshow(canvas[..., ::-1]), plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Matching example - Semi-Dense"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "start = time()\n",
    "mkpts_0, mkpts_1, conf = RDD.match_dense(im0, im1, resize=1024)\n",
    "print(f\"Found {len(mkpts_0)} matches in {time()-start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "canvas = draw_matches(mkpts_0, mkpts_1, im0, im1)\n",
    "plt.figure(figsize=(12,12))\n",
    "plt.imshow(canvas[..., ::-1]), plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Matching example - LightGlue"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "start = time()\n",
    "mkpts_0, mkpts_1, conf = RDD.match_lg(im0, im1, resize=1024)\n",
    "print(f\"Found {len(mkpts_0)} matches in {time()-start:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "canvas = draw_matches(mkpts_0, mkpts_1, im0, im1)\n",
    "plt.figure(figsize=(12,12))\n",
    "plt.imshow(canvas[..., ::-1]), plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Matching example - Using differnt detector + RDD descriptor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "start = time()\n",
    "mkpts_0, mkpts_1, conf = RDD.match_3rd_party(im0, im1, resize=1024, model='aliked')\n",
    "print(f\"Found {len(mkpts_0)} matches in {time()-start:.2f} seconds\")\n",
    "\n",
    "# take a look at folder third_party, RDD/RDD.py and RDD/RDD_helper.py \n",
    "# if you want to configure your own detector"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [
    "KM1KQaj9-oOv"
   ],
   "provenance": [],
   "toc_visible": true
  },
  "kernelspec": {
   "display_name": "RDD",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.18"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}