explore_pattern_detection.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import cv2\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.pyplot import imshow\n",
    "import numpy as np\n",
    "from PIL import Image\n",
    "#from matplotlib.pyplot import figure\n",
    "#plt.figure(figsize = (20,10))\n",
    "plt.rcParams['figure.figsize'] = [10, 10]\n",
    "plt.rcParams['figure.dpi'] = 80\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# Read the original image\n",
    "%matplotlib inline\n",
    "import glob\n",
    "raw_dir='/Users/shamik/Downloads/Chip_images/'\n",
    "fnamelist = [filepath for filepath in glob.glob(raw_dir+'WIN_*.jpg')]\n",
    "for ik in fnamelist:\n",
    "    plt.figure(figsize = (10,10))\n",
    "    img = cv2.imread(ik)\n",
    "    # Display original image\n",
    "    plt.imshow(img)#,aspect='auto')\n",
    "    plt.axis('off')\n",
    "    plt.show()\n",
    "\n",
    "    \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "for ik in fnamelist:\n",
    "    print(ik)\n",
    "    img = cv2.imread(ik)\n",
    "    \n",
    "    plt.figure(figsize = (10,10))\n",
    "    print(\"original image grayscale\")\n",
    "    img_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)\n",
    "    plt.imshow(img_gray, cmap='gray',aspect='auto')\n",
    "    plt.axis('off')\n",
    "    plt.show()\n",
    "\n",
    "    plt.figure(figsize = (10,10))\n",
    "    print(\"cropped image\")\n",
    "    img_crop = img_gray[0:300,:]\n",
    "    plt.imshow(img_crop, cmap='gray')#,aspect='auto')\n",
    "    plt.axis('off')\n",
    "    plt.show()\n",
    "\n",
    "    plt.figure(figsize = (10,10))\n",
    "    print(\"binary threshold\")\n",
    "    ### binary thresholding\n",
    "    ret, img_bin = cv2.threshold(img_crop.copy(), 100, 255, cv2.THRESH_BINARY)\n",
    "    # visualize the binary image\n",
    "    plt.imshow(img_bin, cmap='gray')#,aspect='auto')\n",
    "    plt.axis('off')\n",
    "    plt.show()\n",
    "    \n",
    "    imgin=img_bin\n",
    "    contours, hierarchy = cv2.findContours(imgin, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)\n",
    "    im2 = np.expand_dims(imgin, axis=2).repeat(3, axis=2) \n",
    "\n",
    "    print(\"get contours\")\n",
    "    im2 = cv2.drawContours(im2, contours, -1, (0, 230, 255), 6)\n",
    "    plt.figure(figsize = (10,10))\n",
    "    plt.imshow(im2)\n",
    "    plt.show()\n",
    "\n",
    "\n",
    "    c = max(contours, key = cv2.contourArea)\n",
    "    x,y,w,h = cv2.boundingRect(c)\n",
    "    print(\"get biggest contur bounding box\")\n",
    "    # draw the biggest contour (c) in green\n",
    "    cv2.rectangle(im2,(x,y),(x+w,y+h),(0,255,0),2)\n",
    "    plt.figure(figsize = (10,10))\n",
    "    plt.imshow(im2)\n",
    "    plt.show()\n",
    "\n",
    "    print(\"plot image with box\")\n",
    "    plt.figure(figsize = (10,10))\n",
    "    boximg = cv2.rectangle(img_crop.copy(),(x,y),(x+w,y+h),(0,255,0),2)\n",
    "    plt.imshow(boximg,cmap='gray')\n",
    "    plt.show()\n",
    "\n",
    "    print(\"plot image within box\")\n",
    "    plt.figure(figsize = (10,10))\n",
    "    img_sel = img_crop[y:y+h,x:x+w]\n",
    "    plt.imshow(img_sel,cmap='gray')\n",
    "    plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "refimg = '/Users/shamik/Downloads/Chip_images/WIN_20211220_16_20_46_Pro.jpg'\n",
    "print(\"select template\")\n",
    "\n",
    "img = cv2.imread(refimg)\n",
    "    \n",
    "plt.figure(figsize = (10,10))\n",
    "print(\"original image grayscale\")\n",
    "img_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)\n",
    "plt.imshow(img_gray, cmap='gray',aspect='auto')\n",
    "plt.axis('off')\n",
    "plt.show()\n",
    "\n",
    "plt.figure(figsize = (10,10))\n",
    "print(\"cropped image\")\n",
    "img_crop = img_gray[0:300,:]\n",
    "plt.imshow(img_crop, cmap='gray')#,aspect='auto')\n",
    "plt.axis('off')\n",
    "plt.show()\n",
    "\n",
    "plt.figure(figsize = (10,10))\n",
    "print(\"binary threshold\")\n",
    "### binary thresholding\n",
    "ret, img_bin = cv2.threshold(img_crop.copy(), 100, 255, cv2.THRESH_BINARY)\n",
    "# visualize the binary image\n",
    "plt.imshow(img_bin, cmap='gray')#,aspect='auto')\n",
    "plt.axis('off')\n",
    "plt.show()\n",
    "\n",
    "imgin=img_bin\n",
    "contours, hierarchy = cv2.findContours(imgin, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)\n",
    "im2 = np.expand_dims(imgin, axis=2).repeat(3, axis=2) \n",
    "\n",
    "print(\"get contours\")\n",
    "im2 = cv2.drawContours(im2, contours, -1, (0, 230, 255), 6)\n",
    "plt.figure(figsize = (10,10))\n",
    "plt.imshow(im2)\n",
    "plt.show()\n",
    "\n",
    "\n",
    "c = max(contours, key = cv2.contourArea)\n",
    "x,y,w,h = cv2.boundingRect(c)\n",
    "print(\"get biggest contur bounding box\")\n",
    "# draw the biggest contour (c) in green\n",
    "cv2.rectangle(im2,(x,y),(x+w,y+h),(0,255,0),2)\n",
    "plt.figure(figsize = (10,10))\n",
    "plt.imshow(im2)\n",
    "plt.show()\n",
    "\n",
    "print(\"plot image with box\")\n",
    "plt.figure(figsize = (10,10))\n",
    "boximg = cv2.rectangle(img_crop.copy(),(x,y),(x+w,y+h),(0,255,0),2)\n",
    "plt.imshow(boximg,cmap='gray')\n",
    "plt.show()\n",
    "\n",
    "print(\"plot image within box\")\n",
    "plt.figure(figsize = (10,10))\n",
    "img_sel = img_crop[y:y+h,x:x+w]\n",
    "plt.imshow(img_sel,cmap='gray')\n",
    "plt.show()\n",
    "\n",
    "\n",
    "h1=115\n",
    "w1=125\n",
    "y1=140\n",
    "x1=480\n",
    "img_sig = img_sel[y1:y1+h1,x1:x1+w1]\n",
    "plt.figure(figsize = (10,10))\n",
    "plt.imshow(img_sig,cmap='gray')\n",
    "plt.show( )\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "from skimage.metrics import structural_similarity\n",
    "%matplotlib inline\n",
    "for ik in fnamelist:\n",
    "    print(ik)\n",
    "    img = cv2.imread(ik)\n",
    "    \n",
    "    #plt.figure(figsize = (10,10))\n",
    "    print(\"original image grayscale\")\n",
    "    img_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)\n",
    "    plt.imshow(img_gray, cmap='gray',aspect='auto')\n",
    "    plt.axis('off')\n",
    "    plt.show()\n",
    "    \n",
    "    \n",
    "    \n",
    "\n",
    "\n",
    "    methods = ['cv2.TM_CCOEFF']#, 'cv2.TM_CCOEFF_NORMED', 'cv2.TM_CCORR',\n",
    "                #'cv2.TM_CCORR_NORMED', 'cv2.TM_SQDIFF', 'cv2.TM_SQDIFF_NORMED']\n",
    "    \n",
    "    img_cropn = img_gray\n",
    "    imgc = img_cropn.copy()\n",
    "    template = img_sig.copy()\n",
    "    for meth in methods:\n",
    "        wt, ht = template.shape[::-1]\n",
    "        #img = img2.copy()\n",
    "        method = eval(meth)\n",
    "        # Apply template Matching\n",
    "        res = cv2.matchTemplate(imgc,template,method)\n",
    "        #print(res)\n",
    "        \n",
    "        print(np.max(res))\n",
    "        threshold = 0.99*np.max(res)\n",
    "        loc = np.where( res >= threshold)\n",
    "\n",
    "        for pt in zip(*loc[::-1]):\n",
    "            #print(pt)\n",
    "            cv2.rectangle(imgc, pt, (pt[0] + wt, pt[1] + ht), (0,0,255), 2)\n",
    "            #break\n",
    "\n",
    "\n",
    "        plt.subplot(121)\n",
    "        plt.imshow(res,cmap = 'gray')\n",
    "        plt.title('Matching Result')\n",
    "        plt.xticks([])\n",
    "        plt.yticks([])\n",
    "        plt.subplot(122)\n",
    "        plt.imshow(imgc)#,cmap = 'gray')\n",
    "        plt.title('Detected Point')\n",
    "        plt.xticks([])\n",
    "        plt.yticks([])\n",
    "        plt.suptitle(meth)\n",
    "        plt.show()\n",
    "        \n",
    "        plt.subplot(121)\n",
    "        plt.title(\"image inside the detected box\")\n",
    "        fil_img = img_cropn.copy()[pt[1]:pt[1] + ht, pt[0]:pt[0] + wt]\n",
    "        plt.imshow(fil_img)\n",
    "        #plt.show()\n",
    "        plt.subplot(122)\n",
    "        plt.title(\"flipped image inside the detected box\")\n",
    "        fil_img_flip = np.flip(fil_img)\n",
    "        plt.imshow(fil_img_flip)\n",
    "        plt.show()\n",
    "        \n",
    "        plt.subplot(121)\n",
    "        (score, diff) = structural_similarity(template, fil_img,full=True)\n",
    "        plt.title(\"Image similarity same :\"+ str(score))\n",
    "        plt.imshow(diff)\n",
    "        #plt.show()\n",
    "        plt.subplot(122)\n",
    "        (score, diff) = structural_similarity(template, fil_img_flip,full=True)\n",
    "        plt.title(\"Image similarity flipped :\"+ str(score))\n",
    "        plt.imshow(diff)\n",
    "        plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}