Broiler_Distance_Measure.ipynb

{"cells":[{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":1000,"output_embedded_package_id":"1EVTgxXvV4kRYDtyH1AUGOEFpGzUfINhf"},"executionInfo":{"elapsed":147774,"status":"ok","timestamp":1670430145493,"user":{"displayName":"Mahyar mohtavipour","userId":"09223488037123794171"},"user_tz":-210},"id":"nExg5N4fIA55","outputId":"06062f39-e170-427e-e01e-5b318b391b95"},"outputs":[{"output_type":"display_data","data":{"text/plain":"Output hidden; open in https://colab.research.google.com to view."},"metadata":{}}],"source":["######### System Configuration Output\n","!cat /proc/cpuinfo\n","!nvidia-smi\n","######### System Configuration Output\n","\n","######### Libraries and Directories\n","import os\n","from google.colab import drive\n","drive.mount('/content/drive')\n","os.chdir('/content/drive/MyDrive/Colab Notebooks/ICCKE2022_Challenge')\n","%cd multi-object-tracker\n","!pip install -r requirements.txt\n","!pip install -e \n","from motrackers import CentroidTracker\n","from motrackers.utils import draw_tracks\n","import cv2\n","import numpy as np\n","import matplotlib.pyplot as plt\n","import matplotlib.patches as patches\n","import time\n","from random import randint\n","from google.colab.patches import cv2_imshow\n","os.chdir('/content/drive/MyDrive/Colab Notebooks/ICCKE2022_Challenge/yolo-v4-tf.keras')\n","######## Libraries and Directories\n","\n","######## Loading The Trained Yolov4 Model for Detection\n","class_name_path = '/content/drive/My Drive/Colab Notebooks/ICCKE2022_Challenge/Classes/classes.txt'\n","weight_path = '/content/drive/My Drive/Colab Notebooks/ICCKE2022_Challenge/darknet/yolov4.conv.13'\n","path = '/content/drive/MyDrive/Colab Notebooks/ICCKE2022_Challenge/final_training'\n","from models import Yolov4\n","model = Yolov4(weight_path=weight_path,class_name_path=class_name_path)\n","model.load_model(path)\n","######## Loading The Trained Yolov4 Model for Detection\n","\n","######## Directory for Video Test\n","videoPath = '/content/drive/My Drive/Colab Notebooks/ICCKE2022_Challenge/video_test/1.mp4'\n","######## Directory for Video Test\n","\n","######## Detection Function, Input:Frame , Output: Bounding Boxes with Colors, Score of Detections, Classes of Detections\n","def my_detection(my_frame):\n","  t1 = time.perf_counter()\n","  out,my_colors= model.predict_img(my_frame,plot_img=False)\n","  out_np = out.to_numpy()\n","  bb_xywh = np.append(out_np[:,0:2],  out_np[:,6:8], axis=1)\n","  bb_score = out_np[:,5]\n","  bb_class = out_np[:,0]\n","  bb_xywh = bb_xywh.astype(int)\n","  bb_score = bb_score.astype(int)\n","  bb_class = bb_class.astype(int)\n","  t2 = time.perf_counter()\n","  print('Detection Time Overhead:',t2-t1)\n","  return my_colors, bb_xywh, bb_score, bb_class  \n","######## Detection Function, Input:Frame , Output: Bounding Boxes with Colors, Score of Detections, Classes of Detections\n","\n","######## Setting the Tracker\n","tracker = CentroidTracker(max_lost=5, tracker_output_format='mot_challenge')\n","######## Setting the Tracker\n","\n","######## Capturing the Video Input\n","cap = cv2.VideoCapture(videoPath)\n","length_video = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))\n","success, frame = cap.read()\n","frame=cv2.resize(frame, (480, 480))\n","######## Capturing the Video Input\n","\n","######## Frames list for making output video \n","frames=[]\n","######## Frames list for making output video \n","\n","######## Initializing the Detection and Tracking\n","my_colors,my_bbxywh, my_score, my_class = my_detection(frame)\n","tracks = tracker.update(my_bbxywh, my_score, my_class)\n","######## Initializing the Detection and Tracking\n","\n","######## Data configuration based on Tracking ID, 7-->Total Distance Pixel , 8-->x_center of bbox, 9-->y_center of bbox\n","my_tracks=np.array(tracks)\n","xcentroid, ycentroid = (my_tracks[:,2] + 0.5*my_tracks[:,4]), (my_tracks[:,3] + 0.5*my_tracks[:,5])\n","my_tracks[:,7]= 0\n","my_tracks[:,8]= xcentroid\n","my_tracks[:,9]= ycentroid\n","######## Data configuration based on Tracking ID, 7-->Total Distance Pixel , 8-->x_center of bbox, 9-->y_center of bbox\n","\n","######## Video Detection and Tracking Start\n","for i in range(300): \n","  t1 = time.perf_counter()\n","  print('Frame Number:',i)\n","  ret, frame = cap.read()\n","  frame=cv2.resize(frame, (480, 480))\n","\n","  ######## Storing Previous Tracks Data and Loading New Tracks Data\n","  prev = my_tracks\n","  my_colors,my_bbxywh, my_score, my_class = my_detection(frame)\n","  tracks = tracker.update(my_bbxywh, my_score, my_class)\n","  my_tracks=np.array(tracks)\n","  xcentroid, ycentroid = (my_tracks[:,2] + 0.5*my_tracks[:,4]), (my_tracks[:,3] + 0.5*my_tracks[:,5])\n","  my_tracks[:,8]= xcentroid\n","  my_tracks[:,9]= ycentroid\n","  ######## Storing Previous Tracks Data and Loading New Tracks Data\n","\n","  ######## Data of each Tracking ID\n","  for trk in tracks:\n","    trk_id = trk[1]\n","    xmin = int(trk[2])\n","    ymin = int(trk[3])\n","    width = int(trk[4])\n","    height = int(trk[5])\n","    xcentroid_show, ycentroid_show = int(xmin + 0.5*width), int(ymin + 0.5*height)\n","    ######## Calculating the Cumulative Distance based on Tracking IDs \n","    if (int(trk[1]) in prev[:,1])==True:\n","      index=np.where(my_tracks[:,1]==int(trk[1]))\n","      prev_index=np.where(prev[:,1]==int(trk[1]))\n","      Distance = pow(pow(my_tracks[index,8]-prev[prev_index,8],2)+pow(my_tracks[index,9]-prev[prev_index,9],2),0.5)\n","      ######## Checking the Appearing and Disappearing\n","      if Distance>30:\n","        Distance=0\n","      my_tracks[index,7] = prev[prev_index,7] + Distance\n","      ######## Total Distance Pixel Stores in my_tracks[:,7]\n","      TDP=my_tracks[index,7]\n","      if TDP==-1:\n","        TDP=TDP+1\n","      ######## Displaying the Results\n","      show_TDP=TDP.astype('float64')\n","      show_TDP=np.round(show_TDP,decimals = 1)\n","      cv2.putText(frame, 'TDP='+str(show_TDP[0][0]), (xmin, ymin), cv2.FONT_HERSHEY_SIMPLEX, 0.3, (50, 255, 50), 1)\n","    else:\n","      TDP = 0\n","      cv2.putText(frame, 'TDP='+str(TDP), (xmin, ymin), cv2.FONT_HERSHEY_SIMPLEX, 0.3, (50, 255, 50), 1)\n","    text = \"ID {}\".format(trk_id)\n","    cv2.putText(frame, text, (xcentroid_show - 10, ycentroid_show - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.3, (50, 255, 50), 1)\n","    cv2.circle(frame, (xcentroid_show, ycentroid_show), 2, (0, 255, 0), -1)\n","    cv2.rectangle(frame, (xmin, ymin), (xmin + width, ymin + height), (50, 255, 50), 1)\n","  \n","  ######## Appending the Processed Frames\n","  t2 = time.perf_counter()\n","  print('Total Detection and Tracking Time:',t2-t1)\n","  if (i % 7==0): \n","    cv2_imshow(frame)\n","  frames.append(frame)\n","\n","\n","######## Making the Output Video\n","video = cv2.VideoWriter('Tracking_Result_1.avi', 0, 25, (480, 480))\n","for img in frames: \n","  video.write(img)\n","video.release()\n","######## Making the Output Video\n","\n","######## Video Detection and Tracking Stop\n","\n","\n","\n"]}],"metadata":{"colab":{"provenance":[{"file_id":"1splijlRewdKU24gwhEgzz4f5p2i1i0mc","timestamp":1669848780222}]},"gpuClass":"standard","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":"3.7.4"}},"nbformat":4,"nbformat_minor":0}