update

Author: vvEverett

examples/example_notebooks/mot_YOLOv3 copy.ipynb

@@ -9,13 +9,13 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 8,
+   "execution_count": 1,
    "metadata": {},
    "outputs": [],
    "source": [
     "import numpy as np\n",
     "import cv2 as cv\n",
-    "from motrackers.detectors import nanodet\n",
+    "from motrackers.detectors import Nanodet\n",
     "from motrackers import CentroidTracker, CentroidKF_Tracker, SORT, IOUTracker\n",
     "from motrackers.utils import draw_tracks\n",
     "from nanodet.util import Logger, cfg, load_config, load_model_weight\n",
@@ -24,7 +24,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 9,
+   "execution_count": 2,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -35,21 +35,21 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 10,
+   "execution_count": 3,
    "metadata": {},
    "outputs": [
     {
      "data": {
       "application/vnd.jupyter.widget-view+json": {
-       "model_id": "87a5b23e8af54b2aae7a0234a88ab351",
+       "model_id": "51e2f69a12a44156811e0421ab0e949c",
        "version_major": 2,
        "version_minor": 0
       },
       "text/plain": [
        "Select(description='MOTracker:', options=('CentroidTracker', 'CentroidKF_Tracker', 'SORT', 'IOUTracker'), valu…"
       ]
      },
-     "execution_count": 10,
+     "execution_count": 3,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -67,7 +67,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 11,
+   "execution_count": 4,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -86,7 +86,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 12,
+   "execution_count": 5,
    "metadata": {},
    "outputs": [
     {
@@ -103,26 +103,25 @@
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "\u001b[1m\u001b[35m[root]\u001b[0m\u001b[34m[04-10 20:46:36]\u001b[0m\u001b[32mINFO:\u001b[0m\u001b[37mPress \"Esc\", \"q\" or \"Q\" to exit.\u001b[0m\n",
-      "\u001b[1m\u001b[35m[root]\u001b[0m\u001b[34m[04-10 20:46:36]\u001b[0m\u001b[32mINFO:\u001b[0m\u001b[37mPress \"Esc\", \"q\" or \"Q\" to exit.\u001b[0m\n"
+      "\u001b[1m\u001b[35m[root]\u001b[0m\u001b[34m[04-10 21:59:08]\u001b[0m\u001b[32mINFO:\u001b[0m\u001b[37mPress \"Esc\", \"q\" or \"Q\" to exit.\u001b[0m\n"
      ]
     }
    ],
    "source": [
     "# 导入模型文件\n",
     "local_rank = 0\n",
-    "model = WEIGHTS_PATH\n",
+    "modelpath = WEIGHTS_PATH\n",
     "device = \"cpu:0\"\n",
     "config = CONFIG_FILE_PATH\n",
     "logger = Logger(local_rank, use_tensorboard=False)\n",
     "load_config(cfg, config)\n",
-    "detmodel = nanodet.Nanodet(cfg, model, logger, device)\n",
+    "detmodel = Nanodet(cfg, modelpath, logger, device)\n",
     "logger.log('Press \"Esc\", \"q\" or \"Q\" to exit.')"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 13,
+   "execution_count": 6,
    "metadata": {
     "scrolled": false
    },
@@ -138,12 +137,10 @@
     "            print(\"Cannot read the video feed.\")\n",
     "            break\n",
     "        \n",
-    "        meta, res = detmodel.inference(image)\n",
-    "        bboxes, confidences, class_ids = detmodel.visualize(res[0], meta, cfg.class_names, 0.43)\n",
+    "        meta, res = model.inference(image)\n",
+    "        bboxes,confidences,class_ids,updated_image  = model.visualize(res[0], meta, cfg.class_names, 0.43)\n",
     "        \n",
     "        tracks = tracker.update(bboxes, confidences, class_ids)\n",
-    "        \n",
-    "        updated_image = detmodel.draw_bboxes(image.copy(), bboxes, confidences, class_ids)\n",
     "\n",
     "        updated_image = draw_tracks(updated_image, tracks)\n",
     "\n",
@@ -157,31 +154,39 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 14,
+   "execution_count": 7,
    "metadata": {},
    "outputs": [
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "forward time: 0.146s | decode time: 0.004s | viz time: 0.000s\n"
+      "forward time: 0.156s | decode time: 0.038s | viz time: 0.003s\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "d:\\ProgramData\\Anaconda3\\envs\\ist\\lib\\site-packages\\torch\\functional.py:478: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at  C:\\actions-runner\\_work\\pytorch\\pytorch\\builder\\windows\\pytorch\\aten\\src\\ATen\\native\\TensorShape.cpp:2895.)\n",
+      "  return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]\n"
      ]
     },
     {
-     "ename": "AttributeError",
-     "evalue": "'Nanodet' object has no attribute 'draw_bboxes'",
+     "ename": "ValueError",
+     "evalue": "not enough values to unpack (expected 4, got 3)",
      "output_type": "error",
      "traceback": [
       "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
-      "\u001b[1;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
-      "Cell \u001b[1;32mIn[14], line 1\u001b[0m\n\u001b[1;32m----> 1\u001b[0m main(VIDEO_FILE, model, tracker)\n",
-      "Cell \u001b[1;32mIn[13], line 16\u001b[0m, in \u001b[0;36mmain\u001b[1;34m(video_path, model, tracker)\u001b[0m\n\u001b[0;32m     12\u001b[0m bboxes, confidences, class_ids \u001b[39m=\u001b[39m detmodel\u001b[39m.\u001b[39mvisualize(res[\u001b[39m0\u001b[39m], meta, cfg\u001b[39m.\u001b[39mclass_names, \u001b[39m0.43\u001b[39m)\n\u001b[0;32m     14\u001b[0m tracks \u001b[39m=\u001b[39m tracker\u001b[39m.\u001b[39mupdate(bboxes, confidences, class_ids)\n\u001b[1;32m---> 16\u001b[0m updated_image \u001b[39m=\u001b[39m detmodel\u001b[39m.\u001b[39;49mdraw_bboxes(image\u001b[39m.\u001b[39mcopy(), bboxes, confidences, class_ids)\n\u001b[0;32m     18\u001b[0m updated_image \u001b[39m=\u001b[39m draw_tracks(updated_image, tracks)\n\u001b[0;32m     20\u001b[0m cv\u001b[39m.\u001b[39mimshow(\u001b[39m\"\u001b[39m\u001b[39mimage\u001b[39m\u001b[39m\"\u001b[39m, updated_image)\n",
-      "\u001b[1;31mAttributeError\u001b[0m: 'Nanodet' object has no attribute 'draw_bboxes'"
+      "\u001b[1;31mValueError\u001b[0m                                Traceback (most recent call last)",
+      "Cell \u001b[1;32mIn[7], line 1\u001b[0m\n\u001b[1;32m----> 1\u001b[0m main(VIDEO_FILE, detmodel, tracker)\n",
+      "Cell \u001b[1;32mIn[6], line 12\u001b[0m, in \u001b[0;36mmain\u001b[1;34m(video_path, model, tracker)\u001b[0m\n\u001b[0;32m      9\u001b[0m     \u001b[39mbreak\u001b[39;00m\n\u001b[0;32m     11\u001b[0m meta, res \u001b[39m=\u001b[39m model\u001b[39m.\u001b[39minference(image)\n\u001b[1;32m---> 12\u001b[0m bboxes,confidences,class_ids,updated_image  \u001b[39m=\u001b[39m model\u001b[39m.\u001b[39mvisualize(res[\u001b[39m0\u001b[39m], meta, cfg\u001b[39m.\u001b[39mclass_names, \u001b[39m0.43\u001b[39m)\n\u001b[0;32m     14\u001b[0m tracks \u001b[39m=\u001b[39m tracker\u001b[39m.\u001b[39mupdate(bboxes, confidences, class_ids)\n\u001b[0;32m     16\u001b[0m updated_image \u001b[39m=\u001b[39m draw_tracks(updated_image, tracks)\n",
+      "\u001b[1;31mValueError\u001b[0m: not enough values to unpack (expected 4, got 3)"
      ]
     }
    ],
    "source": [
-    "main(VIDEO_FILE, model, tracker)"
+    "main(VIDEO_FILE, detmodel, tracker)"
    ]
   },
   {

main.py

@@ -1,24 +1,24 @@
 import numpy as np
 import cv2 as cv
-from motrackers.detectors import nanodet
+from motrackers.detectors import Nanodet
 from motrackers import CentroidTracker, CentroidKF_Tracker, SORT, IOUTracker
 from motrackers.utils import draw_tracks
 from nanodet.util import Logger, cfg, load_config, load_model_weight
 
-VIDEO_FILE = "22.avi"
-WEIGHTS_PATH = 'weight/LiquidV4.pth'
-CONFIG_FILE_PATH = 'config/LiquidDetect416.yml'
+VIDEO_FILE = r"D:\shijue\LiquidDrop\22.avi"
+WEIGHTS_PATH = r'D:\shijue\multi-object-tracker\weight\LiquidV4.pth'
+CONFIG_FILE_PATH = r'D:\shijue\multi-object-tracker\config\LiquidDetect416.yml'
 
 tracker = CentroidTracker(max_lost=0, tracker_output_format='mot_challenge')
 
 # 导入模型文件
 local_rank = 0
-model = WEIGHTS_PATH
+modelpath = WEIGHTS_PATH
 device = "cpu:0"
 config = CONFIG_FILE_PATH
 logger = Logger(local_rank, use_tensorboard=False)
 load_config(cfg, config)
-model = nanodet(cfg, model, logger, device)
+detmodel = Nanodet(cfg, modelpath, logger, device)
 logger.log('Press "Esc", "q" or "Q" to exit.')
 
 def main(video_path, model, tracker):
@@ -32,11 +32,9 @@ def main(video_path, model, tracker):
             break
 
         meta, res = model.inference(image)
-        bboxes, confidences, class_ids = model.visualize(res[0], meta, cfg.class_names, 0.43)
+        bboxes,confidences,class_ids,updated_image  = model.visualize(res[0], meta, cfg.class_names, 0.43)
 
         tracks = tracker.update(bboxes, confidences, class_ids)
-        
-        updated_image = model.draw_bboxes(image.copy(), bboxes, confidences, class_ids)
 
         updated_image = draw_tracks(updated_image, tracks)
 
@@ -49,4 +47,4 @@ def main(video_path, model, tracker):
 
 
 
-main(VIDEO_FILE, model, tracker)
+main(VIDEO_FILE, detmodel, tracker)

motrackers/detectors/__init__.py

@@ -1,3 +1,4 @@
 from motrackers.detectors.tf import TF_SSDMobileNetV2
 from motrackers.detectors.caffe import Caffe_SSDMobileNet
 from motrackers.detectors.yolo import YOLOv3
+from motrackers.detectors.nanodet import Nanodet

motrackers/detectors/nanodet.py

@@ -63,7 +63,19 @@ def visualize(self, dets, meta, class_names, score_thres, wait=0):
         self.class_names = class_names
         bboxes , confidences , class_ids = infotrans(all_box)
         print("viz time: {:.3f}s".format(time.time() - time1))
-        return bboxes , confidences , class_ids
+        
+        image = meta["raw_img"][0].copy()
+        for bb, conf, cid in zip(bboxes, confidences, class_ids):
+            bbox_colors = {key: np.random.randint(0, 255, size=(3,)).tolist() for key in self.class_names.keys()}
+            clr = [int(c) for c in bbox_colors[cid]]
+            cv2.rectangle(image, (bb[0], bb[1]), (bb[0] + bb[2], bb[1] + bb[3]), clr, 2)
+            label = "{}:{:.4f}".format(self.class_names[cid], conf)
+            (label_width, label_height), baseLine = cv2.getTextSize(label, cv2.FONT_HERSHEY_SIMPLEX, 0.5, 2)
+            y_label = max(bb[1], label_height)
+            cv2.rectangle(image, (bb[0], y_label - label_height), (bb[0] + label_width, y_label + baseLine),
+                            (255, 255, 255), cv2.FILLED)
+            cv2.putText(image, label, (bb[0], y_label), cv2.FONT_HERSHEY_SIMPLEX, 0.5, clr, 2)
+        return bboxes , confidences , class_ids , image
 
     def draw_bboxes(self, image, bboxes, confidences, class_ids):
         """

setup_nanodet.py

@@ -0,0 +1,27 @@
+#!/usr/bin/env python
+from setuptools import find_packages, setup
+
+from nanodet import __author__, __author_email__, __docs__, __homepage__, __version__
+
+if __name__ == "__main__":
+    setup(
+        name="nanodet",
+        version=__version__,
+        description=__docs__,
+        url=__homepage__,
+        author=__author__,
+        author_email=__author_email__,
+        keywords="deep learning",
+        packages=find_packages(exclude=("config", "tools", "demo")),
+        classifiers=[
+            "Development Status :: Beta",
+            "License :: OSI Approved :: Apache Software License",
+            "Operating System :: OS Independent",
+            "Programming Language :: Python :: 3.5",
+            "Programming Language :: Python :: 3.6",
+            "Programming Language :: Python :: 3.7",
+            "Programming Language :: Python :: 3.8",
+        ],
+        license="Apache License 2.0",
+        zip_safe=False,
+    )