添加文档和整理

01_dataset.ipynb

@@ -13,7 +13,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 25,
+   "execution_count": 1,
    "metadata": {
     "pycharm": {
      "name": "#%%\n"
@@ -40,7 +40,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 26,
+   "execution_count": 2,
    "metadata": {
     "pycharm": {
      "name": "#%%\n"
@@ -49,19 +49,7 @@
    "outputs": [],
    "source": [
     "img_path = r\"data/dataset_old/img/yangeng.bmp\"\n",
-    "label_path = r\"data/dataset_old/label/yangeng.bmp\""
+    "label_path = r\"data/dataset_old/label/yangeng.bmp\"\n",
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 27,
-   "metadata": {
-    "pycharm": {
-     "name": "#%%\n"
-    }
-   },
-   "outputs": [],
-   "source": [
     "# 读取图片和色彩空间转换\n",
     "img = cv2.imread(img_path)\n",
     "label_img = cv2.imread(label_path)\n",

02_classification.ipynb

@@ -7,13 +7,11 @@
      "name": "#%% md\n"
     }
    },
-   "source": [
+   "source": []
-    "# 模型的训练"
-   ]
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": 5,
    "metadata": {
     "pycharm": {
      "name": "#%%\n"
@@ -28,110 +26,139 @@
     "from utils import read_labeled_img"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "pycharm": {
-     "name": "#%% md\n"
-    }
-   },
-   "source": [
-    "## 读取数据与构建数据集"
-   ]
-  },
   {
    "cell_type": "code",
-   "execution_count": 7,
+   "execution_count": null,
-   "metadata": {
-    "pycharm": {
-     "name": "#%%\n"
-    }
-   },
    "outputs": [],
    "source": [
-    "data_dir = \"data/dataset\"\n",
+    "train_from_existed = True # 是否从现有数据训练，如果是的话，那就从dataset_file训练，否则就用data_dir里头的数据\n",
-    "color_dict = {(0, 0, 255): \"yangeng\", (255, 0, 0): 'beijing'}\n",
+    "data_dir = \"data/dataset\" # 数据集，文件夹下必须包含`img`和`label`两个文件夹，放置相同文件名的图片和label\n",
-    "label_index = {\"yangeng\": 1, \"beijing\": 0}\n",
+    "dataset_file = \"data/dataset/dataset_2022-07-20_10-04.mat\"\n",
-    "dataset = read_labeled_img(data_dir, color_dict=color_dict, is_ps_color_space=False)\n",
-    "rus = RandomUnderSampler(random_state=0)\n",
-    "x_list, y_list = np.concatenate([v for k, v in dataset.items()], axis=0).tolist(), \\\n",
-    "                 np.concatenate([np.ones((v.shape[0],)) * label_index[k] for k, v in dataset.items()], axis=0).tolist()\n",
     "\n",
-    "x_resampled, y_resampled = rus.fit_resample(x_list, y_list)\n",
+    "color_dict = {(0, 0, 255): \"yangeng\", (255, 0, 0): 'beijing'} # 颜色对应的类别\n",
-    "dataset = {\"inside\": np.array(x_resampled)}"
+    "label_index = {\"yangeng\": 1, \"beijing\": 0} # 类别对应的序号\n",
-   ]
+    "show_samples = False # 是否展示样本\n",
+    "\n",
+    "# 定义一些训练量\n",
+    "threshold = 5  # 正样本周围多大范围内的还算是正样本\n",
+    "node_num = 20  # 如果使用ELM作为分类器物，有多少的节点\n",
+    "negative_sample_num = None  # None或者一个数字，对应生成的负样本数量"
+   ],
+   "metadata": {
+    "collapsed": false,
+    "pycharm": {
+     "name": "#%%\n"
+    }
+   }
   },
   {
    "cell_type": "markdown",
-   "metadata": {
-    "pycharm": {
-     "name": "#%% md\n"
-    }
-   },
    "source": [
-    "## 模型训练"
+    "## 读取数据"
-   ]
+   ],
+   "metadata": {
+    "collapsed": false
+   }
   },
   {
    "cell_type": "code",
-   "execution_count": 8,
+   "execution_count": 10,
-   "metadata": {
-    "pycharm": {
-     "name": "#%%\n"
-    }
-   },
-   "outputs": [],
-   "source": [
-    "# 定义一些常量\n",
-    "threshold = 5\n",
-    "node_num = 20\n",
-    "negative_sample_num = None  # None或者一个数字\n",
-    "world_boundary = np.array([0, 0, 0, 255, 255, 255])\n",
-    "# 对数据进行预处理\n",
-    "x = np.concatenate([v for k, v in dataset.items()], axis=0)\n",
-    "negative_sample_num = int(x.shape[0] * 1.2) if negative_sample_num is None else negative_sample_num\n"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 2,
-   "metadata": {
-    "pycharm": {
-     "name": "#%%\n"
-    }
-   },
-   "outputs": [],
-   "source": [
-    "model = AnonymousColorDetector()"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 3,
    "outputs": [
     {
-     "name": "stdout",
+     "ename": "FileNotFoundError",
-     "output_type": "stream",
+     "evalue": "[Errno 2] No such file or directory: 'data/dataset/label'",
-     "text": [
+     "output_type": "error",
-      "              precision    recall  f1-score   support\n",
+     "traceback": [
-      "\n",
+      "\u001B[0;31m---------------------------------------------------------------------------\u001B[0m",
-      "         0.0       0.99      0.99      0.99     26314\n",
+      "\u001B[0;31mFileNotFoundError\u001B[0m                         Traceback (most recent call last)",
-      "         1.0       0.99      0.99      0.99     24492\n",
+      "\u001B[0;32m/var/folders/wh/kr5c3dr12834pfk3j7yqnrq40000gn/T/ipykernel_30867/1942905945.py\u001B[0m in \u001B[0;36m<module>\u001B[0;34m\u001B[0m\n\u001B[0;32m----> 1\u001B[0;31m \u001B[0mdataset\u001B[0m \u001B[0;34m=\u001B[0m \u001B[0mread_labeled_img\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0mdata_dir\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0mcolor_dict\u001B[0m\u001B[0;34m=\u001B[0m\u001B[0mcolor_dict\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0mis_ps_color_space\u001B[0m\u001B[0;34m=\u001B[0m\u001B[0;32mFalse\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n\u001B[0m\u001B[1;32m      2\u001B[0m \u001B[0;32mif\u001B[0m \u001B[0mshow_samples\u001B[0m\u001B[0;34m:\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n\u001B[1;32m      3\u001B[0m     \u001B[0;32mfrom\u001B[0m \u001B[0mutils\u001B[0m \u001B[0;32mimport\u001B[0m \u001B[0mlab_scatter\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n\u001B[1;32m      4\u001B[0m     \u001B[0mlab_scatter\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0mdataset\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0mclass_max_num\u001B[0m\u001B[0;34m=\u001B[0m\u001B[0;36m30000\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0mis_3d\u001B[0m\u001B[0;34m=\u001B[0m\u001B[0;32mTrue\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0mis_ps_color_space\u001B[0m\u001B[0;34m=\u001B[0m\u001B[0;32mFalse\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n",
-      "\n",
+      "\u001B[0;32m~/PycharmProjects/tobacco_color/utils.py\u001B[0m in \u001B[0;36mread_labeled_img\u001B[0;34m(dataset_dir, color_dict, ext, is_ps_color_space)\u001B[0m\n\u001B[1;32m     37\u001B[0m     \u001B[0;34m:\u001B[0m\u001B[0;32mreturn\u001B[0m\u001B[0;34m:\u001B[0m \u001B[0m字典形式的数据集\u001B[0m\u001B[0;34m{\u001B[0m\u001B[0mlabel\u001B[0m\u001B[0;34m:\u001B[0m \u001B[0mvector\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0mn\u001B[0m \u001B[0mx\u001B[0m \u001B[0;36m3\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m}\u001B[0m\u001B[0;34m,\u001B[0m\u001B[0mvector为lab色彩空间\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n\u001B[1;32m     38\u001B[0m     \"\"\"\n\u001B[0;32m---> 39\u001B[0;31m     img_names = [img_name for img_name in os.listdir(os.path.join(dataset_dir, 'label'))\n\u001B[0m\u001B[1;32m     40\u001B[0m                  if img_name.endswith(ext)]\n\u001B[1;32m     41\u001B[0m     \u001B[0mtotal_dataset\u001B[0m \u001B[0;34m=\u001B[0m \u001B[0mMergeDict\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n",
-      "    accuracy                           0.99     50806\n",
+      "\u001B[0;31mFileNotFoundError\u001B[0m: [Errno 2] No such file or directory: 'data/dataset/label'"
-      "   macro avg       0.99      0.99      0.99     50806\n",
-      "weighted avg       0.99      0.99      0.99     50806\n",
-      "\n"
      ]
     }
    ],
    "source": [
-    "# model.fit(x, world_boundary, threshold, negative_sample_size=negative_sample_num, train_size=0.7,\n",
+    "dataset = read_labeled_img(data_dir, color_dict=color_dict, is_ps_color_space=False)\n",
-    "#           is_save_dataset=True, model_selection='dt')\n",
+    "if show_samples:\n",
-    "data = scipy.io.loadmat('data/dataset/dataset_2022-07-20_10-04.mat')\n",
+    "    from utils import lab_scatter\n",
-    "x, y = data['x'], data['y'].ravel()\n",
+    "    lab_scatter(dataset, class_max_num=30000, is_3d=True, is_ps_color_space=False)"
-    "model.fit(x, y=y, is_generate_negative=False, model_selection='dt')\n",
+   ],
-    "model.save()"
+   "metadata": {
+    "collapsed": false,
+    "pycharm": {
+     "name": "#%%\n"
+    }
+   }
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "## 数据平衡化"
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "outputs": [
+    {
+     "ename": "NameError",
+     "evalue": "name 'dataset' is not defined",
+     "output_type": "error",
+     "traceback": [
+      "\u001B[0;31m---------------------------------------------------------------------------\u001B[0m",
+      "\u001B[0;31mNameError\u001B[0m                                 Traceback (most recent call last)",
+      "\u001B[0;32m/var/folders/wh/kr5c3dr12834pfk3j7yqnrq40000gn/T/ipykernel_30867/603974095.py\u001B[0m in \u001B[0;36m<module>\u001B[0;34m\u001B[0m\n\u001B[1;32m      1\u001B[0m \u001B[0mrus\u001B[0m \u001B[0;34m=\u001B[0m \u001B[0mRandomUnderSampler\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0mrandom_state\u001B[0m\u001B[0;34m=\u001B[0m\u001B[0;36m0\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n\u001B[0;32m----> 2\u001B[0;31m \u001B[0mx_list\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0my_list\u001B[0m \u001B[0;34m=\u001B[0m \u001B[0mnp\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mconcatenate\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0;34m[\u001B[0m\u001B[0mv\u001B[0m \u001B[0;32mfor\u001B[0m \u001B[0mk\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0mv\u001B[0m \u001B[0;32min\u001B[0m \u001B[0mdataset\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mitems\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m]\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0maxis\u001B[0m\u001B[0;34m=\u001B[0m\u001B[0;36m0\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mtolist\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m,\u001B[0m\u001B[0;31m \u001B[0m\u001B[0;31m\\\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n\u001B[0m\u001B[1;32m      3\u001B[0m                  \u001B[0mnp\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mconcatenate\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0;34m[\u001B[0m\u001B[0mnp\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mones\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0mv\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mshape\u001B[0m\u001B[0;34m[\u001B[0m\u001B[0;36m0\u001B[0m\u001B[0;34m]\u001B[0m\u001B[0;34m,\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m)\u001B[0m \u001B[0;34m*\u001B[0m \u001B[0mlabel_index\u001B[0m\u001B[0;34m[\u001B[0m\u001B[0mk\u001B[0m\u001B[0;34m]\u001B[0m \u001B[0;32mfor\u001B[0m \u001B[0mk\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0mv\u001B[0m \u001B[0;32min\u001B[0m \u001B[0mdataset\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mitems\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m]\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0maxis\u001B[0m\u001B[0;34m=\u001B[0m\u001B[0;36m0\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mtolist\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n\u001B[1;32m      4\u001B[0m \u001B[0mx_resampled\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0my_resampled\u001B[0m \u001B[0;34m=\u001B[0m \u001B[0mrus\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mfit_resample\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0mx_list\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0my_list\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n\u001B[1;32m      5\u001B[0m \u001B[0mdataset\u001B[0m \u001B[0;34m=\u001B[0m \u001B[0;34m{\u001B[0m\u001B[0;34m\"inside\"\u001B[0m\u001B[0;34m:\u001B[0m \u001B[0mnp\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0marray\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0mx_resampled\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m}\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n",
+      "\u001B[0;31mNameError\u001B[0m: name 'dataset' is not defined"
+     ]
+    }
+   ],
+   "source": [
+    "rus = RandomUnderSampler(random_state=0)\n",
+    "x_list, y_list = np.concatenate([v for k, v in dataset.items()], axis=0).tolist(), \\\n",
+    "                 np.concatenate([np.ones((v.shape[0],)) * label_index[k] for k, v in dataset.items()], axis=0).tolist()\n",
+    "x_resampled, y_resampled = rus.fit_resample(x_list, y_list)\n",
+    "dataset = {\"inside\": np.array(x_resampled)}"
+   ],
+   "metadata": {
+    "collapsed": false,
+    "pycharm": {
+     "name": "#%%\n"
+    }
+   }
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "## 模型训练"
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "outputs": [
+    {
+     "ename": "NameError",
+     "evalue": "name 'dataset' is not defined",
+     "output_type": "error",
+     "traceback": [
+      "\u001B[0;31m---------------------------------------------------------------------------\u001B[0m",
+      "\u001B[0;31mNameError\u001B[0m                                 Traceback (most recent call last)",
+      "\u001B[0;32m/var/folders/wh/kr5c3dr12834pfk3j7yqnrq40000gn/T/ipykernel_30867/1828483636.py\u001B[0m in \u001B[0;36m<module>\u001B[0;34m\u001B[0m\n\u001B[1;32m      1\u001B[0m \u001B[0;31m# 对数据进行预处理\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n\u001B[0;32m----> 2\u001B[0;31m \u001B[0mx\u001B[0m \u001B[0;34m=\u001B[0m \u001B[0mnp\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mconcatenate\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0;34m[\u001B[0m\u001B[0mv\u001B[0m \u001B[0;32mfor\u001B[0m \u001B[0mk\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0mv\u001B[0m \u001B[0;32min\u001B[0m \u001B[0mdataset\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mitems\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m]\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0maxis\u001B[0m\u001B[0;34m=\u001B[0m\u001B[0;36m0\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n\u001B[0m\u001B[1;32m      3\u001B[0m \u001B[0mnegative_sample_num\u001B[0m \u001B[0;34m=\u001B[0m \u001B[0mint\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0mx\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mshape\u001B[0m\u001B[0;34m[\u001B[0m\u001B[0;36m0\u001B[0m\u001B[0;34m]\u001B[0m \u001B[0;34m*\u001B[0m \u001B[0;36m1.2\u001B[0m\u001B[0;34m)\u001B[0m \u001B[0;32mif\u001B[0m \u001B[0mnegative_sample_num\u001B[0m \u001B[0;32mis\u001B[0m \u001B[0;32mNone\u001B[0m \u001B[0;32melse\u001B[0m \u001B[0mnegative_sample_num\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n\u001B[1;32m      4\u001B[0m \u001B[0mmodel\u001B[0m \u001B[0;34m=\u001B[0m \u001B[0mAnonymousColorDetector\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n",
+      "\u001B[0;31mNameError\u001B[0m: name 'dataset' is not defined"
+     ]
+    }
+   ],
+   "source": [
+    "# 对数据进行预处理\n",
+    "x = np.concatenate([v for k, v in dataset.items()], axis=0)\n",
+    "negative_sample_num = int(x.shape[0] * 1.2) if negative_sample_num is None else negative_sample_num\n",
+    "model = AnonymousColorDetector()"
    ],
    "metadata": {
     "collapsed": false,
@@ -142,9 +169,31 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 14,
-   "outputs": [],
+   "outputs": [
-   "source": [],
+    {
+     "ename": "NameError",
+     "evalue": "name 'train_from_existed' is not defined",
+     "output_type": "error",
+     "traceback": [
+      "\u001B[0;31m---------------------------------------------------------------------------\u001B[0m",
+      "\u001B[0;31mNameError\u001B[0m                                 Traceback (most recent call last)",
+      "\u001B[0;32m/var/folders/wh/kr5c3dr12834pfk3j7yqnrq40000gn/T/ipykernel_30867/103680055.py\u001B[0m in \u001B[0;36m<module>\u001B[0;34m\u001B[0m\n\u001B[0;32m----> 1\u001B[0;31m \u001B[0;32mif\u001B[0m \u001B[0mtrain_from_existed\u001B[0m\u001B[0;34m:\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n\u001B[0m\u001B[1;32m      2\u001B[0m     \u001B[0mdata\u001B[0m \u001B[0;34m=\u001B[0m \u001B[0mscipy\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mio\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mloadmat\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0mdataset_file\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n\u001B[1;32m      3\u001B[0m     \u001B[0mx\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0my\u001B[0m \u001B[0;34m=\u001B[0m \u001B[0mdata\u001B[0m\u001B[0;34m[\u001B[0m\u001B[0;34m'x'\u001B[0m\u001B[0;34m]\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0mdata\u001B[0m\u001B[0;34m[\u001B[0m\u001B[0;34m'y'\u001B[0m\u001B[0;34m]\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mravel\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n\u001B[1;32m      4\u001B[0m     \u001B[0mmodel\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0mfit\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0mx\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0my\u001B[0m\u001B[0;34m=\u001B[0m\u001B[0my\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0mis_generate_negative\u001B[0m\u001B[0;34m=\u001B[0m\u001B[0;32mFalse\u001B[0m\u001B[0;34m,\u001B[0m \u001B[0mmodel_selection\u001B[0m\u001B[0;34m=\u001B[0m\u001B[0;34m'dt'\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n\u001B[1;32m      5\u001B[0m \u001B[0mmodel\u001B[0m\u001B[0;34m.\u001B[0m\u001B[0msave\u001B[0m\u001B[0;34m(\u001B[0m\u001B[0;34m)\u001B[0m\u001B[0;34m\u001B[0m\u001B[0;34m\u001B[0m\u001B[0m\n",
+      "\u001B[0;31mNameError\u001B[0m: name 'train_from_existed' is not defined"
+     ]
+    }
+   ],
+   "source": [
+    "if train_from_existed:\n",
+    "    data = scipy.io.loadmat(dataset_file)\n",
+    "    x, y = data['x'], data['y'].ravel()\n",
+    "    model.fit(x, y=y, is_generate_negative=False, model_selection='dt')\n",
+    "else:\n",
+    "    world_boundary = np.array([0, 0, 0, 255, 255, 255])\n",
+    "    model.fit(x, world_boundary, threshold, negative_sample_size=negative_sample_num, train_size=0.7,\n",
+    "          is_save_dataset=True, model_selection='dt')\n",
+    "model.save()"
+   ],
    "metadata": {
     "collapsed": false,
     "pycharm": {

README.md

@@ -2,4 +2,68 @@
 
 2022年7月18日开始开发的项目，使用彩色相机进行烟梗颜色的识别。
 
-项目中的数据放置于[链接](https://macrosolid-my.sharepoint.com/:f:/g/personal/feijinti_miaow_fun/ElIAerke0rVMjcSyrbjetAYBpide1s-rPEY9JDZWrrhiUA?e=hrzOfa)
+## 如何进行模型训练和部署？
+1. 项目当中需要包含`data`和`models`这两个文件夹，请下载到当前文件夹下,这是链接：[data](https://macrosolid-my.sharepoint.com/personal/feijinti_miaow_fun/_layouts/15/onedrive.aspx?id=%2Fpersonal%2Ffeijinti%5Fmiaow%5Ffun%2FDocuments%2FPycharmProjects%2Ftobacco%5Fcolor%2Fdata&ga=1), [models](https://macrosolid-my.sharepoint.com/:f:/g/personal/feijinti_miaow_fun/EiyBjWEX90JGn8S-e5Kh7N8B1GWvfvDcNbpleWDTwkDm1w?e=wyL4EF)
+2. 使用[01_dataset.ipynb](./01_dataset.ipynb) 进行数据集的分析文件格式需要设置为这种形式：
+    ```text
+    dataset
+    ├── label
+    │   ├── img1.bmp
+    │   └── ...
+    └── img
+        ├── img1.bmp
+        └── ...
+    ```
+3. 使用[02_classification.ipynb](./02_classification.ipynb)进行训练
+4. 使用[03_data_update.ipynb](02_classification.ipynb)进行数据的更新与添加
+5. 使用`main_test.py`文件进行读图测试
+6. **部署**，复制`utils.py`、`models.py`、`main.py`、`models`、`config.py`这5个文件或文件夹，运行main.py来提供预测服务。
+
+
+## 训练的原理
+为了应对工业环境当中负样本少的特点，我们结合颜色有限空间的特性对我们的训练过程进行了优化，核心的优化方式在于制造负样本
+
+### 负样本是怎么造出来的？
+
+ 我们对于一个给定的色彩空间进行随机的生成一些数据，然后判断它是否是给定的正样本附近，如果是在附近，那么我们就把这些点看作是正样本，如果离得比较远，那么就会被当作是负样本。
+
+### 训练的结果
+
+这样子进行训练，模型就会被约束在我们给定的样本范围内，就像你看到的这样。
+
+![image-20220721153751824](https://raw.githubusercontent.com/Karllzy/imagebed/main/img/image-20220721153751824.png)
+
+在这里，绿色就是目标的色彩范围，橙色的和蓝色则表明了模型的判定范围，模型认为蓝色的区域就是烟梗，而橙色的区域就不是烟梗。
+
+可以看到，蓝色区域与绿色区域是高度重叠的，并且蓝色比绿色区域要大一些的，这正是我们想要的效果。这表明模型对于烟梗的颜色有适度的宽容，允许色彩有一定的偏差，但大体上是要达到烟梗颜色范围内的。
+
+这样的好处在于，即使出现了新的杂质，只要这些杂质的色彩不在模型的宽容范围内（蓝色范围内），那么都会被判定为杂质。
+
+## 预测过程的后处理（异色问题）
+
+### 问题的发现
+
+![image-20220721154731187](https://raw.githubusercontent.com/Karllzy/imagebed/main/img/image-20220721154731187.png)
+
+在摄影过程中，由于相机、镜头和拍摄物体多方面的原因会出现色散边的现象，就像上图这样本来应该黄色的烟梗，边缘却变成了绿色或紫色的。
+
+这是由于不同波长的光折射率不同，到达成像单元的位置会出现细小的偏差，而我们的成像单元又比较的细小，举例来说，这可能使得物体上同样的点发出的红光到了1号像素，而发出的绿光本来应该也射到1号像素却射到了相邻的2号像素，这就导致色彩不对了。
+
+根据资料，一般的解决方案是对于不同波长的光进行折射率补偿，使用抗色散镜头。
+
+## 镜头的影响
+
+我们现有的镜头包括广角和窄角两个，这两个镜头有着不同的成像效果，如下图所示。
+
+| 视角     |                           广角镜头                           |                           窄角镜头                           |
+| -------- | :----------------------------------------------------------: | :----------------------------------------------------------: |
+| 普通视角 | ![image-20220721155919349](https://raw.githubusercontent.com/Karllzy/imagebed/main/img/image-20220721155919349.png) | ![image-20220721155848160](https://raw.githubusercontent.com/Karllzy/imagebed/main/img/image-20220721155848160.png) |
+
+但是由于条件有限，我们这里就只能用算法的形式硬抗这些误差了。
+
+
+
+## 模型的更新
+
+### 如何应对新的目标物？
+