{"cells":[{"cell_type":"code","source":["from google.colab import drive\n","drive.mount('/content/drive')"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"1VkPIQMBmJMO","executionInfo":{"status":"ok","timestamp":1654700494173,"user_tz":-60,"elapsed":3080,"user":{"displayName":"Adejumo Daniel","userId":"02925977078148845759"}},"outputId":"1e3cd91c-ca69-486b-b182-d2f31583b645"},"execution_count":1,"outputs":[{"output_type":"stream","name":"stdout","text":["Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"]}]},{"cell_type":"code","source":["%cd ./drive/MyDrive/Python/Machine Learning/Computer Vision/darknet-COCO-object_detection"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"izkEuuuPmTZf","executionInfo":{"status":"ok","timestamp":1654700494174,"user_tz":-60,"elapsed":11,"user":{"displayName":"Adejumo Daniel","userId":"02925977078148845759"}},"outputId":"daf99ba8-1ed2-4935-e2b9-3481fef9584a"},"execution_count":2,"outputs":[{"output_type":"stream","name":"stdout","text":["/content/drive/MyDrive/Python/Machine Learning/Computer Vision/darknet-COCO-object_detection\n"]}]},{"cell_type":"code","source":["!pip install filterpy --quiet"],"metadata":{"id":"qXFwvyxqmXDr","executionInfo":{"status":"ok","timestamp":1654700498924,"user_tz":-60,"elapsed":4757,"user":{"displayName":"Adejumo Daniel","userId":"02925977078148845759"}}},"execution_count":3,"outputs":[]},{"cell_type":"code","source":["!pip install lap --quiet"],"metadata":{"id":"zqK3-Fn2oRsc","executionInfo":{"status":"ok","timestamp":1654700503070,"user_tz":-60,"elapsed":4165,"user":{"displayName":"Adejumo Daniel","userId":"02925977078148845759"}}},"execution_count":4,"outputs":[]},{"cell_type":"code","execution_count":5,"metadata":{"id":"kHwKuAkPlviV","executionInfo":{"status":"ok","timestamp":1654700504310,"user_tz":-60,"elapsed":1248,"user":{"displayName":"Adejumo Daniel","userId":"02925977078148845759"}}},"outputs":[],"source":["from models import *\n","from utils import *\n","\n","import os, sys, time, datetime, random\n","import torch\n","from torch.utils.data import DataLoader\n","from torchvision import datasets, transforms\n","from torch.autograd import Variable\n","\n","import matplotlib.pyplot as plt\n","import matplotlib.patches as patches\n","from PIL import Image"]},{"cell_type":"code","execution_count":6,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"N5uZwVlClvie","executionInfo":{"status":"ok","timestamp":1654700508098,"user_tz":-60,"elapsed":3795,"user":{"displayName":"Adejumo Daniel","userId":"02925977078148845759"}},"outputId":"3a3e75b1-3379-4e79-f418-0b8a48ffb62f"},"outputs":[{"output_type":"stream","name":"stderr","text":["/usr/local/lib/python3.7/dist-packages/torch/nn/_reduction.py:42: UserWarning: size_average and reduce args will be deprecated, please use reduction='mean' instead.\n","  warnings.warn(warning.format(ret))\n"]}],"source":["config_path='config/yolov3.cfg'\n","weights_path='config/yolov3.weights'\n","class_path='config/coco.names'\n","img_size=416\n","conf_thres=0.8\n","nms_thres=0.4\n","\n","# Load model and weights\n","model = Darknet(config_path, img_size=img_size)\n","model.load_weights(weights_path)\n","model.cuda()\n","model.eval()\n","classes = utils.load_classes(class_path)\n","Tensor = torch.cuda.FloatTensor"]},{"cell_type":"code","execution_count":7,"metadata":{"id":"n4NNQSOYlvij","executionInfo":{"status":"ok","timestamp":1654700508099,"user_tz":-60,"elapsed":9,"user":{"displayName":"Adejumo Daniel","userId":"02925977078148845759"}}},"outputs":[],"source":["def detect_image(img):\n","    # scale and pad image\n","    ratio = min(img_size/img.size[0], img_size/img.size[1])\n","    imw = round(img.size[0] * ratio)\n","    imh = round(img.size[1] * ratio)\n","    img_transforms = transforms.Compose([ transforms.Resize((imh, imw)),\n","         transforms.Pad((max(int((imh-imw)/2),0), max(int((imw-imh)/2),0), max(int((imh-imw)/2),0), max(int((imw-imh)/2),0)),\n","                        (128,128,128)),\n","         transforms.ToTensor(),\n","         ])\n","    # convert image to Tensor\n","    image_tensor = img_transforms(img).float()\n","    image_tensor = image_tensor.unsqueeze_(0)\n","    input_img = Variable(image_tensor.type(Tensor))\n","    # run inference on the model and get detections\n","    with torch.no_grad():\n","        detections = model(input_img)\n","        detections = utils.non_max_suppression(detections, 80, conf_thres, nms_thres)\n","    return detections[0]"]},{"cell_type":"code","execution_count":8,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"7VUHsE2-lvik","executionInfo":{"status":"ok","timestamp":1654700521379,"user_tz":-60,"elapsed":13287,"user":{"displayName":"Adejumo Daniel","userId":"02925977078148845759"}},"outputId":"6144a350-24e9-4a7c-95c5-96bb66b824e0"},"outputs":[{"output_type":"stream","name":"stdout","text":["Populating the interactive namespace from numpy and matplotlib\n"]},{"output_type":"stream","name":"stderr","text":["/usr/local/lib/python3.7/dist-packages/IPython/core/magics/pylab.py:161: UserWarning: pylab import has clobbered these variables: ['random']\n","`%matplotlib` prevents importing * from pylab and numpy\n","  \"\\n`%matplotlib` prevents importing * from pylab and numpy\"\n"]},{"output_type":"stream","name":"stdout","text":["Video size 1280 720\n"]},{"output_type":"stream","name":"stderr","text":["/content/drive/MyDrive/Python/Machine Learning/Computer Vision/darknet-COCO-object_detection/sort.py:38: NumbaWarning: \n","Compilation is falling back to object mode WITH looplifting enabled because Function \"iou\" failed type inference due to: non-precise type pyobject\n","During: typing of argument at /content/drive/MyDrive/Python/Machine Learning/Computer Vision/darknet-COCO-object_detection/sort.py (43)\n","\n","File \"sort.py\", line 43:\n","def iou(bb_test,bb_gt):\n","    <source elided>\n","  \"\"\"\n","  xx1 = np.maximum(bb_test[0], bb_gt[0])\n","  ^\n","\n","  @jit\n","/usr/local/lib/python3.7/dist-packages/numba/core/object_mode_passes.py:178: NumbaWarning: Function \"iou\" was compiled in object mode without forceobj=True.\n","\n","File \"sort.py\", line 39:\n","@jit\n","def iou(bb_test,bb_gt):\n","^\n","\n","  state.func_ir.loc))\n","/usr/local/lib/python3.7/dist-packages/numba/core/object_mode_passes.py:188: NumbaDeprecationWarning: \n","Fall-back from the nopython compilation path to the object mode compilation path has been detected, this is deprecated behaviour.\n","\n","For more information visit https://numba.pydata.org/numba-doc/latest/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit\n","\n","File \"sort.py\", line 39:\n","@jit\n","def iou(bb_test,bb_gt):\n","^\n","\n","  state.func_ir.loc))\n"]}],"source":["videopath = './videos/HorseRacing.mp4'\n","\n","%pylab inline \n","import cv2\n","from IPython.display import clear_output\n","\n","cmap = plt.get_cmap('tab20b')\n","colors = [cmap(i)[:3] for i in np.linspace(0, 1, 20)]\n","\n","# initialize Sort object and video capture\n","from sort import *\n","vid = cv2.VideoCapture(videopath)\n","mot_tracker = Sort()\n","\n","fourcc = cv2.VideoWriter_fourcc(*'XVID')\n","ret,frame=vid.read()\n","vw = frame.shape[1]\n","vh = frame.shape[0]\n","print (\"Video size\", vw,vh)\n","outvideo = cv2.VideoWriter(videopath.replace(\".mp4\", \"-det.mp4\"),fourcc,20.0,(vw,vh))\n","\n","# while(True):\n","for ii in range(40):\n","    ret, frame = vid.read()\n","    if not ret:\n","      print(\"Done Procesing Video\")\n","      break\n","    frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)\n","    pilimg = Image.fromarray(frame)\n","    detections = detect_image(pilimg)\n","\n","    img = np.array(pilimg)\n","    pad_x = max(img.shape[0] - img.shape[1], 0) * (img_size / max(img.shape))\n","    pad_y = max(img.shape[1] - img.shape[0], 0) * (img_size / max(img.shape))\n","    unpad_h = img_size - pad_y\n","    unpad_w = img_size - pad_x\n","    if detections is not None:\n","        tracked_objects = mot_tracker.update(detections.cpu())\n","\n","        unique_labels = detections[:, -1].cpu().unique()\n","        n_cls_preds = len(unique_labels)\n","        for x1, y1, x2, y2, obj_id, cls_pred in tracked_objects:\n","            box_h = int(((y2 - y1) / unpad_h) * img.shape[0])\n","            box_w = int(((x2 - x1) / unpad_w) * img.shape[1])\n","            y1 = int(((y1 - pad_y // 2) / unpad_h) * img.shape[0])\n","            x1 = int(((x1 - pad_x // 2) / unpad_w) * img.shape[1])\n","\n","            color = colors[int(obj_id) % len(colors)]\n","            color = [i * 255 for i in color]\n","            cls = classes[int(cls_pred)]\n","            cv2.rectangle(frame, (x1, y1), (x1+box_w, y1+box_h), color, 4)\n","            cv2.rectangle(frame, (x1, y1-35), (x1+len(cls)*19+60, y1), color, -1)\n","            cv2.putText(frame, cls + \"-\" + str(int(obj_id)), (x1, y1 - 10), cv2.FONT_HERSHEY_SIMPLEX, 1, (255,255,255), 3)\n","\n","    outvideo.write(frame)\n","\n","outvideo.release()"]},{"cell_type":"code","source":["from pathlib import Path\n","from IPython import display as ipythondisplay\n","import base64\n","\n","def show_videos(video_path='', prefix=''):\n","  html = []\n","  for mp4 in Path(video_path).glob(f\"{prefix}*.mp4\"):\n","      video_b64 = base64.b64encode(mp4.read_bytes())\n","      html.append('''<video alt=\"{}\" autoplay \n","                    loop controls style=\"height: 400px;\">\n","                    <source src=\"data:video/mp4;base64,{}\" type=\"video/mp4\" />\n","                </video>'''.format(mp4, video_b64.decode('ascii')))\n","      break\n","  ipythondisplay.display(ipythondisplay.HTML(data=\"<br>\".join(html)))"],"metadata":{"id":"Xx6d_F3VstfA","executionInfo":{"status":"ok","timestamp":1654700521380,"user_tz":-60,"elapsed":19,"user":{"displayName":"Adejumo Daniel","userId":"02925977078148845759"}}},"execution_count":9,"outputs":[]},{"cell_type":"code","source":["video_b64 = base64.b64encode(Path(videopath.replace(\".mp4\", \"-det.mp4\")).read_bytes())\n","html = '''<video alt=\"{}\" autoplay \n","              loop controls style=\"height: 400px;\">\n","              <source src=\"data:video/mp4;base64,{}\" type=\"video/mp4\" />\n","          </video>'''.format(Path(videopath), video_b64.decode('ascii'))\n","ipythondisplay.display(ipythondisplay.HTML(data=html))     "],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":421,"output_embedded_package_id":"1KE6a6Jf_qBrnIGEjOY8GYXagvaaGt84D"},"id":"K3VrKNb3yUbH","executionInfo":{"status":"ok","timestamp":1654700524974,"user_tz":-60,"elapsed":3611,"user":{"displayName":"Adejumo Daniel","userId":"02925977078148845759"}},"outputId":"92ea1435-9e17-4167-c094-dd1e380b200f"},"execution_count":10,"outputs":[{"output_type":"display_data","data":{"text/plain":"Output hidden; open in https://colab.research.google.com to view."},"metadata":{}}]}],"metadata":{"kernelspec":{"display_name":"Python 3","name":"python3"},"language_info":{"name":"python"},"colab":{"name":"PyTorch_Object_Tracking.ipynb","provenance":[],"collapsed_sections":[]},"accelerator":"GPU"},"nbformat":4,"nbformat_minor":0}