pybaseutils 2.4.34__tar.gz → 2.5.0__tar.gz

{pybaseutils-2.4.34 → pybaseutils-2.5.0}/PKG-INFO

@@ -1,13 +1,19 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.4
 Name: pybaseutils
-Version: 2.4.34
+Version: 2.5.0
 Summary: pybaseutils
 Home-page: https://github.com/PanJinquan/base-utils
 Author: PanJinquan
 Author-email: 390737991@qq.com
 License: MIT
-Platform: UNKNOWN
 License-File: LICENCE
+Dynamic: author
+Dynamic: author-email
+Dynamic: description
+Dynamic: home-page
+Dynamic: license
+Dynamic: license-file
+Dynamic: summary
 
 base-utils
 ==========
@@ -142,5 +148,3 @@ base_utils(C++)
        DEBUG_IMSHOW("image", image);
        return 0;
    }
-
-

{pybaseutils-2.4.34 → pybaseutils-2.5.0}/pybaseutils/__init__.py

@@ -8,4 +8,4 @@
 # --------------------------------------------------------
 """
 
-__version__ = '2.4.34'
+__version__ = '2.5.0'

{pybaseutils-2.4.34 → pybaseutils-2.5.0}/pybaseutils/audio/audio_utils.py

@@ -19,7 +19,6 @@ from ffmpy import FFmpeg
 import matplotlib.pyplot as plt
 from scipy.fft import fft
 from IPython.display import Audio, display
-from playsound import playsound
 from scipy.io import wavfile
 
 import numpy as np
@@ -267,6 +266,8 @@ def sound_audio(audio, sr=16000):
 
 def sound_audio_file(audio_file, jupyter=False):
     """
+    sudo apt-get install libasound2-dev
+    pip install simpleaudio==1.0.4
     :param audio_file:
     :param jupyter:
     :return:
@@ -274,7 +275,10 @@ def sound_audio_file(audio_file, jupyter=False):
     if jupyter:
         display(Audio(audio_file))
     else:
-        playsound(audio_file)
+        import simpleaudio as sa
+        wave_obj = sa.WaveObject.from_wave_file(audio_file)
+        play_obj = wave_obj.play() # 播放
+        play_obj.wait_done()  # 等待播放结束（可选，但推荐）
 
 
 def sound_audio_buffer(audio_buffer, tmp_audio="./tmp.wav"):
@@ -470,8 +474,8 @@ if __name__ == '__main__':
     video_out = "../../data/video/test-video-result.mp4"
     audio_file = "../../data/audio/bus_chinese.wav"
 
-    test_time(audio_file)
-    # playsound(audio_file)
+    # test_time(audio_file)
+    sound_audio_file(audio_file,jupyter=False)
     # playsound(audio_file)
     # display_time_domain(audio_file)
     # display_freq_domain(audio_file)

{pybaseutils-2.4.34 → pybaseutils-2.5.0}/pybaseutils/base64_utils.py

@@ -11,13 +11,16 @@ import sys
 import os
 
 import PIL.Image as Image
+from PIL import ImageFile
 import cv2
 import base64
 import numpy as np
 from typing import Any
+from io import BytesIO
 
 IMG_PREFIX = "image/jpg"  # 图片base64字符串前缀
 precision = 6  # 小数点精度
+ImageFile.LOAD_TRUNCATED_IMAGES = True  # 允许加载截断的图像
 
 
 def isbase64(data: str, prefix=IMG_PREFIX):
@@ -36,10 +39,14 @@ def base642image(bs64, prefix=IMG_PREFIX, use_rgb=False) -> np.ndarray:
     if prefix == bs64[0:len(prefix)]:
         bs64 = bs64[len(prefix):]
     bs64 = bytes(bs64, 'utf-8')
-    image = base64.b64decode(bs64)
-    image = np.frombuffer(image, np.uint8)
-    image = cv2.imdecode(image, flags=cv2.IMREAD_UNCHANGED)
+    imgbf = base64.b64decode(bs64)
+    # image = np.frombuffer(imgbf, np.uint8)
+    # image = cv2.imdecode(image, flags=cv2.IMREAD_UNCHANGED)
     # image = cv2.imdecode(image, cv2.IMREAD_COLOR)
+    # if image is None:
+    image = Image.open(BytesIO(imgbf)) # 修复截断的图像
+    image = np.array(image)
+    image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
     if use_rgb:
         image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
     return image

{pybaseutils-2.4.34 → pybaseutils-2.5.0}/pybaseutils/color_utils.py

@@ -10,6 +10,7 @@
 import cv2
 import matplotlib.pyplot as plt
 import numpy as np
+from collections import defaultdict, Counter
 
 VOC_NAMES = ["aeroplane", "bicycle", "bird", "boat", "bottle", "bus", "car", "cat",
              "chair", "cow", "diningtable", "dog", "horse", "motorbike", "person",
@@ -28,6 +29,30 @@ COCO_NAMES = ['background', 'person', 'bicycle', 'car', 'motorcycle', 'airplane'
               'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors',
               'teddy bear', 'hair drier', 'toothbrush']
 
+color_table = [(0, 0, 0), (0, 255, 0), (255, 0, 0), (0, 0, 255), (0, 255, 255), (255, 0, 255), (255, 255, 0),
+               (128, 0, 0), (0, 128, 0), (128, 128, 0),
+               (0, 0, 128), (128, 0, 128), (0, 128, 128), (128, 128, 128),
+               (64, 0, 0), (192, 0, 0), (64, 128, 0), (192, 128, 0),
+               (64, 0, 128), (192, 0, 128), (64, 128, 128), (192, 128, 128),
+               (0, 64, 0), (128, 64, 0), (0, 192, 0), (128, 192, 0), (0, 64, 128)] * 100
+
+
+class ColorLabel():
+    def __init__(self):
+        self.color = dict()
+        self.cnums = len(color_table)
+
+    def get_color(self, label: str):
+        """
+        :param label: str or list
+        :return:
+        """
+        h = str(label)
+        if h not in self.color:
+            i = len(self.color) + 1
+            self.color[h] = color_table[i % self.cnums]
+        return self.color[h]
+
 
 def get_colormap(data_type="custom"):
     if data_type == 'pascal':

{pybaseutils-2.4.34 → pybaseutils-2.5.0}/pybaseutils/dataloader/parser_labelme.py

@@ -537,6 +537,34 @@ def LabelMeDatasets(filename=None,
     return datasets
 
 
+def get_class_counts(datasets, save=""):
+    """
+    统计每个类别的样本数量
+    :param datasets: List[str] or ConcatDataset
+    :param save: 保存路径
+    :return: 每个类别的样本数量
+    """
+    from pybaseutils import dict_utils
+    if isinstance(datasets, list):
+        datasets = LabelMeDatasets(filename=None,
+                                  data_root=None,
+                                  anno_dir=datasets,
+                                  image_dir=None,
+                                  class_name=[],
+                                  check=False,
+                                  phase='val',
+                                  shuffle=False)
+    counts = {}
+    for i in tqdm(range(len(datasets))):
+        info = datasets.__getitem__(i)
+        for n in info["names"]:
+            counts[n] = counts.get(n, 0) + 1
+    counts = dict_utils.dict_sort(counts,use_key=False)
+    if save:
+        json_utils.save_json(os.path.join(save, "counts.json"), counts)
+    return counts
+
+
 def parser_labelme(anno_file, class_dict={}, size=()):
     """
     :param annotation:  labelme标注的数据
@@ -633,4 +661,10 @@ def example_for_segment():
 
 if __name__ == "__main__":
     # example_for_segment()
-    example_for_keypoints()
+    # example_for_keypoints()
+    datasets = [
+        '/home/PKing/edudata/dataset/AICQ/dataset/data-2026-10-06/images',
+        '/home/PKing/edudata/dataset/AICQ/dataset/data-2026-09-30/images',
+    ]
+    counts = get_class_counts(datasets)
+    print(counts)

{pybaseutils-2.4.34 → pybaseutils-2.5.0}/pybaseutils/dataloader/parser_voc.py

@@ -119,7 +119,7 @@ class VOCDataset(Dataset):
             class_set = labels.reshape(-1).tolist() + class_set
             class_set = list(set(class_set))
             if ignore_empty and (len(boxes) == 0 or len(labels) == 0):
-                print("empty annotation:{}".format(annotation_file))
+                # print("empty annotation:{}".format(annotation_file))
                 continue
             dst_ids.append(image_id)
         class_set = sorted(class_set)

{pybaseutils-2.4.34 → pybaseutils-2.5.0}/pybaseutils/file_utils.py

@@ -148,6 +148,25 @@ def get_data_hash(data: np.ndarray):
     return hash
 
 
+def get_hash(image: np.ndarray = None, texts=[]) -> str:
+    """根据image和texts数据生成唯一ID(hash)"""
+    if texts is None: texts = []
+    # 使用更快的哈希算法（如果安全性要求不高）
+    hasher = hashlib.sha256()  # 或 hashlib.md5() for speed
+    # 文本处理
+    if texts:
+        texts = sorted([t.strip() for t in texts if t and t.strip()])
+        if texts: hasher.update('|'.join(texts).encode('utf-8'))
+    # 图像处理
+    if image is not None and isinstance(image, np.ndarray):
+        if not image.flags.c_contiguous:  image = np.ascontiguousarray(image)  # 确保连续性（视图操作）
+        # 使用memoryview避免拷贝
+        hasher.update(memoryview(image))
+    # 如果完全没有数据，返回空哈希
+    v = hasher.hexdigest() if hasher.digest_size > 0 else hashlib.sha256(b'empty').hexdigest()
+    return v
+
+
 def replace_elements(items, src, dst, ignore=True):
     """
     将列表中，值为src改为dst
@@ -264,6 +283,18 @@ def save_json(json_file, json_data):
 write_json_path = save_json
 
 
+def save_jsonl(jsonl_file, data_list: list):
+    """
+    data_list转为jsonl保持
+    :param jsonl_file:
+    :param data_list:
+    :return:
+    """
+    with open(jsonl_file, 'w', encoding='utf-8') as f:
+        for item in data_list:
+            f.write(json.dumps(item, ensure_ascii=False) + '\n')
+
+
 def load_json_files(files: list, max_workers=8):
     """
     根据url读取文件

{pybaseutils-2.4.34 → pybaseutils-2.5.0}/pybaseutils/font_style/__init__.py

@@ -19,6 +19,7 @@ FONT_ROOT = [
     os.path.expanduser("~/.fonts"),  # 用户字体目录
     BASE_FONT,  # 用户字体目录
     "/usr/share/fonts",
+    "./data/fonts",
     "/usr/share/fonts/truetype",
     "/usr/share/fonts/opentype",
 ]

{pybaseutils-2.4.34 → pybaseutils-2.5.0}/pybaseutils/font_utils.py

@@ -243,6 +243,30 @@ def re_example():
     print(out)
 
 
+def align_display(s, align='C', width=20, pchar=" "):
+    """
+    对齐显示字符串
+    print("[{}] 下载进度".format(align_display(item, 'C')))
+    :param s: 要显示的字符串
+    :param align: 对齐方式，'L' 左对齐，'C' 居中对齐，'R' 右对齐
+    :param width: 目标显示宽度
+    :param pchar: 对齐字符，默认空格
+    :return: 对齐后的字符串
+    """
+    w = sum(2 if ord(c) > 127 else 1 for c in s)  # 计算字符串的显示宽度（中文等宽字符算2，英文算1）
+    total_pad = max(width - w, 0)
+    if align == 'L':
+        L = 0
+        R = total_pad
+    elif align == 'R':
+        L = total_pad
+        R = 0
+    else:  # 默认居中对齐
+        L = total_pad // 2
+        R = total_pad - L
+    return pchar * L + s + pchar * R
+
+
 if __name__ == "__main__":
     draw_font_example()
     # get_system_fonts()

{pybaseutils-2.4.34 → pybaseutils-2.5.0}/pybaseutils/image_utils.py

@@ -43,6 +43,8 @@ color_map = [(0, 0, 0), (56, 56, 255), (151, 157, 255), (31, 112, 255), (29, 178
              (168, 153, 44), (255, 194, 0), (147, 69, 52), (255, 115, 100),
              (236, 24, 0), (255, 56, 132), (133, 0, 82), (255, 56, 203), (200, 149, 255), (199, 55, 255)] * 100
 
+color_label = color_utils.ColorLabel()
+
 root = os.path.dirname(__file__)
 coco_skeleton = [[1, 3], [1, 0], [2, 4], [2, 0], [0, 5], [0, 6], [5, 7], [7, 9], [6, 8],
                  [8, 10], [5, 11], [6, 12], [11, 12], [11, 13], [13, 15], [12, 14], [14, 16]]
@@ -1124,13 +1126,14 @@ def draw_image_bboxes_text(image, boxes, boxes_name, color=(), thickness=0, font
     return image
 
 
-def draw_image_boxes_texts(image, boxes, texts, color=(), thickness=0, fontScale=0, alpha=0, drawType="custom",
-                           top=True):
+def draw_image_boxes_texts(image, boxes, texts, color=(), thickness=0, fontScale=0, alpha=0,
+                           drawType="custom", color_type="class", top=True):
     """
     :param image:
     :param boxes:
     :param texts:
     :param color:
+    :param kColor:
     :param thickness:
     :param fontScale:
     :param drawType:
@@ -1138,8 +1141,10 @@ def draw_image_boxes_texts(image, boxes, texts, color=(), thickness=0, fontScale
     :return:
     """
     for i, (bbox, text) in enumerate(zip(boxes, texts)):
-        bbox = [int(b) for b in bbox]
-        c = color if color else color_table[(i + 1) % 20]
+        if color_type == "class":
+            c = color if color else color_label.get_color(text)  # 相同类别相同颜色
+        else:
+            c = color if color else color_table[(i + 1) % len(color_table)]  # 每个实例不同颜色
         draw_image_bbox_text(image, bbox, text, color=c, thickness=thickness, fontScale=fontScale, alpha=alpha,
                              drawType=drawType, top=top)
     return image
@@ -1161,16 +1166,14 @@ def draw_image_bboxes_labels_text(image, boxes, labels, boxes_name=None, color=N
 
 def draw_image_boxes_labels_texts(image, boxes, labels, texts, color=None, thickness=2, fontScale=0.8,
                                   drawType="custom", top=True, colors=None, color_type="class"):
-    if isinstance(labels, np.ndarray):
-        labels = labels.reshape(-1).tolist()
+    if isinstance(labels, np.ndarray): labels = labels.reshape(-1).tolist()
     colors = [(0, 0, 0)] + colors if colors else color_table
-    for i, (label, bbox, name) in enumerate(zip(labels, boxes, texts)):
-        bbox = [int(b) for b in bbox]
+    for i, (label, bbox, text) in enumerate(zip(labels, boxes, texts)):
         if color_type == "class":
-            c = color if color else colors[(int(label) + 1) % 20]  # 相同类别相同颜色
+            c = color if color else color_label.get_color(label)  # 相同类别相同颜色
         else:
-            c = color if color else colors[(i + 1) % 20]  # 每个实例不同颜色
-        image = draw_image_bbox_text(image, bbox, str(name), color=c, thickness=thickness, fontScale=fontScale,
+            c = color if color else colors[(i + 1) % len(colors)]  # 每个实例不同颜色
+        image = draw_image_bbox_text(image, bbox, str(text), color=c, thickness=thickness, fontScale=fontScale,
                                      drawType=drawType, top=top)
     return image
 
@@ -1253,28 +1256,34 @@ def show_image_rects_texts(title, image, rects, texts, color=None, thickness=0,
     return image
 
 
-def draw_image_bboxes_labels(image, boxes, labels, class_name=None, color=None, thickness=0, fontScale=0,
-                             drawType="custom", top=True):
+def draw_image_boxes_labels(image, boxes, labels, class_name=None, color=None, kColor=True, thickness=0, fontScale=0,
+                            alpha=0, drawType="custom", top=True):
     """
-    :param image:
-    :param boxes:  [[x1,y1,x2,y2],[x1,y1,x2,y2]]
-    :param labels:
+    绘制框和标签
+    :param image: 图像
+    :param boxes: 框坐标
+    :param labels: 标签index
+    :param class_name: 类别名称
+    :param color:
+    :param kColor: True 相同类别颜色相同，每个类别颜色不同; False 每个实例颜色不同
+    :param thickness:
+    :param fontScale:
+    :param alpha:
+    :param drawType:
+    :param top:
     :return:
     """
     if isinstance(labels, np.ndarray): labels = labels.astype(np.int32).reshape(-1).tolist()
-    for label, bbox in zip(labels, boxes):
-        bbox = [int(b) for b in bbox]
-        c = color
-        if isinstance(label, numbers.Number) and class_name:
-            c = color_table[(int(label) + 1) % 20]
-            label = class_name[int(label)]
-        if not c: c = color_table[1]
+    if class_name: labels = [class_name[i] for i in labels]
+    for i, (label, bbox) in enumerate(zip(labels, boxes)):
+        c = color if color else color_label.get_color(label)  # 相同类别相同颜色
+        if not kColor: c = color_table[(i + 1) % len(color_table)]
         image = draw_image_bbox_text(image, bbox, str(label), color=c, thickness=thickness, fontScale=fontScale,
-                                     drawType=drawType, top=top)
+                                     alpha=alpha, drawType=drawType, top=top)
     return image
 
 
-draw_image_boxes_labels = draw_image_bboxes_labels
+draw_image_bboxes_labels = draw_image_boxes_labels
 
 
 def draw_image_rects_labels(image, rects, labels, class_name=None, color=None, thickness=0, fontScale=0):
@@ -1467,25 +1476,22 @@ def draw_text(image, point, text, color=(255, 0, 0), fontScale=0, thickness=0, d
     :param image:
     :param point:
     :param text:
-    :param drawType: custom or simple
+    :param drawType: en/ch/simple
     :return:
     """
     point = (int(point[0]), int(point[1]))
     thickness, fontScale = get_linesize(max(image.shape), thickness=thickness, fontScale=fontScale)
-    fontFace = cv2.FONT_HERSHEY_SIMPLEX
-    # fontFace=cv2.FONT_HERSHEY_SIMPLEX
+    fontFace = cv2.FONT_HERSHEY_COMPLEX
     if drawType == "custom" or drawType == "en":
-        text_size, baseline = cv2.getTextSize(str(text), fontFace, fontScale, abs(thickness))
-        text_loc = (point[0], point[1] + text_size[1])
-        cv2.rectangle(image, (text_loc[0] - 2 // 2, text_loc[1] - 2 - baseline),
-                      (text_loc[0] + text_size[0], text_loc[1] + text_size[1]), color=color, thickness=-1)
-        # draw score value
-        cv2.putText(image, str(text), (text_loc[0], text_loc[1] + baseline), fontFace, fontScale, (255, 255, 255),
-                    abs(thickness), 2)
+        bbox = [point[0], point[1]]
+        image = draw_bbox_text_en(image, bbox, text, color=color, thickness=thickness, fontScale=fontScale)
+    elif drawType == "chinese" or drawType == "ch":
+        bbox = [point[0], point[1]]
+        image = draw_bbox_text_ch(image, bbox, text, color=color, thickness=thickness, fontScale=fontScale)
     elif drawType == "simple":
         cv2.putText(image, str(text), (point[0], point[1]), fontFace, fontScale, color=color, thickness=abs(thickness))
-    if drawType == "chinese" or drawType == "ch":
-        cv2_putText(image, str(text), point, color=color, fontScale=fontScale, thickness=abs(thickness))
+    else:
+        cv2.putText(image, str(text), (point[0], point[1]), fontFace, fontScale, color=color, thickness=abs(thickness))
     return image
 
 
@@ -1499,38 +1505,62 @@ def draw_image_bbox_text(image, bbox, text, color, thickness=2, fontScale=0.8, a
     :param top:
     :return:
     """
+    bbox = [int(b) for b in bbox]
     text = str(text)
     if alpha > 0:
         bgimg = draw_image_bbox_text(image.copy(), bbox, text=text, color=color, thickness=thickness,
                                      fontScale=fontScale, drawType=drawType, top=top, alpha=0)
         image[:] = cv2.addWeighted(bgimg, 1 - alpha, image, alpha, 0)
         return image
-    thickness, fontScale = get_linesize(max(image.shape), thickness=thickness, fontScale=fontScale)
-    # text_loc = (bbox[0], bbox[1]) if top else (bbox[0], bbox[3])
-    if not text: drawType = "simple"
-    text_size, baseline = cv2.getTextSize(str(text), cv2.FONT_HERSHEY_SIMPLEX, fontScale, abs(thickness))
-    if top:
-        # text_loc = (bbox[0], bbox[1] + text_size[1])  # 在左上角下方绘制文字
+    if drawType == "ch" or drawType == "chinese":
+        image = draw_bbox_text_ch(image, bbox, text, color, thickness=thickness, fontScale=fontScale)
+    elif drawType == "simple":
+        text_wh, baseline = cv2.getTextSize(str(text), cv2.FONT_HERSHEY_COMPLEX, fontScale, abs(thickness))
         text_loc = (bbox[0], bbox[1] - baseline // 2)  # 在左上角上方绘制文字
-    else:
-        text_loc = (bbox[0], bbox[3] + text_size[1] + baseline // 2)
-    if drawType == "chinese" or drawType == "ch":
-        if not top: text_loc = (text_loc[0], text_loc[1] - baseline)
-        cv2.rectangle(image, (bbox[0], bbox[1]), (bbox[2], bbox[3]), color, thickness)
-        cv2_putText(image, str(text), text_loc, color=color, fontScale=fontScale, thickness=abs(thickness))
-    elif drawType == "simple" or drawType == "en":
         cv2.rectangle(image, (bbox[0], bbox[1]), (bbox[2], bbox[3]), color, thickness, 8, 0)
-        cv2.putText(image, str(text), text_loc, cv2.FONT_HERSHEY_SIMPLEX, fontScale, color, abs(thickness))
-    elif drawType == "custom":
-        cv2.rectangle(image, (bbox[0], bbox[1]), (bbox[2], bbox[3]), color, thickness)
-        bg1 = (text_loc[0], text_loc[1] - text_size[1] - baseline // 2)
-        bg2 = (text_loc[0] + text_size[0], text_loc[1] + baseline // 2)  # 底纹等于字体的宽度
-        # bg2 = (max(bbox[2], text_loc[0] + text_size[0]), text_loc[1] + baseline // 2) # 底纹等于box的宽度
-        cv2.rectangle(image, bg1, bg2, color, thickness)  # 先绘制框，再填充
-        cv2.rectangle(image, bg1, bg2, color, -1)
-        # draw score value
-        cv2.putText(image, str(text), (text_loc[0], text_loc[1]), cv2.FONT_HERSHEY_SIMPLEX, fontScale,
-                    (255, 255, 255), abs(thickness))
+        cv2.putText(image, str(text), text_loc, cv2.FONT_HERSHEY_COMPLEX, fontScale, color, abs(thickness))
+    else:  # custom ,en
+        image = draw_bbox_text_en(image, bbox, text, color, thickness=thickness, fontScale=fontScale)
+    return image
+
+
+def draw_bbox_text_ch(image, bbox, text, color, thickness=2, fontScale=0.8):
+    thickness, fontScale = get_linesize(max(image.shape), thickness=thickness, fontScale=fontScale)
+    text_wh, baseline = cv2.getTextSize(str(text), cv2.FONT_HERSHEY_COMPLEX, fontScale, abs(thickness))
+    baseline = baseline + 1
+    text_wh = (text_wh[0], int(text_wh[1] * 0.8))  # 与draw_bbox_text_en区别，底纹缩小80%
+    text_loc = (bbox[0], bbox[1] - baseline // 2)  # 在左上角上方绘制文字
+    text_bgy = text_loc[1] - text_wh[1] - baseline // 2
+    if text_bgy < 0:  # 如果文字超出图片范围，切换到下方绘制
+        text_loc = (bbox[0], bbox[1] + text_wh[1] + baseline // 2)  # 在左上角下方绘制文字
+    bg1 = (text_loc[0], text_loc[1] - text_wh[1] - baseline // 2)
+    bg2 = (text_loc[0] + text_wh[0], text_loc[1] + baseline // 2)  # 底纹等于字体的宽度
+    # bg2 = (max(bbox[2], text_loc[0] + text_wh[0]), text_loc[1] + baseline // 2) # 底纹等于box的宽度
+    cv2.rectangle(image, bg1, bg2, color, thickness)  # 先绘制框，避免断裂
+    cv2.rectangle(image, bg1, bg2, color, -1)  # 绘制底纹
+    if len(bbox) == 4: cv2.rectangle(image, (bbox[0], bbox[1]), (bbox[2], bbox[3]), color, thickness)  # 绘制bbox
+    text_loc = (text_loc[0], int(text_loc[1] * 1.0)) # 调整字体位置
+    cv2_putText(image, str(text), text_loc, color=(255, 255, 255), fontScale=fontScale, thickness=abs(thickness))
+    return image
+
+
+def draw_bbox_text_en(image, bbox, text, color, thickness=2, fontScale=0.8):
+    thickness, fontScale = get_linesize(max(image.shape), thickness=thickness, fontScale=fontScale)
+    text_wh, baseline = cv2.getTextSize(str(text), cv2.FONT_HERSHEY_SIMPLEX, fontScale, abs(thickness))
+    baseline = baseline + 1
+    text_loc = (bbox[0], bbox[1] - baseline // 2)  # 在左上角上方绘制文字
+    text_bgy = text_loc[1] - text_wh[1] - baseline // 2
+    if text_bgy < 0:  # 如果文字超出图片范围，切换到下方绘制
+        text_loc = (bbox[0], bbox[1] + text_wh[1] + baseline // 2)  # 在左上角下方绘制文字
+    bg1 = (text_loc[0], text_loc[1] - text_wh[1] - baseline // 2)
+    bg2 = (text_loc[0] + text_wh[0], text_loc[1] + baseline // 2)  # 底纹等于字体的宽度
+    # bg2 = (max(bbox[2], text_loc[0] + text_wh[0]), text_loc[1] + baseline // 2) # 底纹等于box的宽度
+    cv2.rectangle(image, bg1, bg2, color, thickness)  # 先绘制框，避免断裂
+    cv2.rectangle(image, bg1, bg2, color, -1)  # 绘制底纹
+    if len(bbox) == 4: cv2.rectangle(image, (bbox[0], bbox[1]), (bbox[2], bbox[3]), color, thickness)  # 绘制bbox
+    text_loc = (text_loc[0], int(text_loc[1] * 0.95)) # 调整字体位置
+    cv2.putText(image, str(text), (text_loc[0], text_loc[1]), fontFace=cv2.FONT_HERSHEY_SIMPLEX,
+                fontScale=fontScale, color=(255, 255, 255), thickness=abs(thickness))
     return image
 
 
@@ -1543,14 +1573,13 @@ def draw_text_line(image, point, text_line: str, bg_color=(255, 0, 0), thickness
     :return:
     """
     thickness, fontScale = get_linesize(max(image.shape), thickness=thickness, fontScale=fontScale)
-    fontFace = cv2.FONT_HERSHEY_SIMPLEX
-    # fontFace=cv2.FONT_HERSHEY_SIMPLEX
+    fontFace = cv2.FONT_HERSHEY_COMPLEX
     text_line = text_line.split("\n")
-    # text_size, baseline = cv2.getTextSize(str(text_line), fontFace, fontScale, thickness)
-    text_size, baseline = cv2.getTextSize(str(text_line), fontFace, fontScale, thickness)
+    # text_wh, baseline = cv2.getTextSize(str(text_line), fontFace, fontScale, thickness)
+    text_wh, baseline = cv2.getTextSize(str(text_line), fontFace, fontScale, thickness)
     for i, text in enumerate(text_line):
         if text:
-            draw_point = (point[0], point[1] + (text_size[1] + 2 + baseline) * i)
+            draw_point = (point[0], point[1] + (text_wh[1] + 2 + baseline) * i)
             image = draw_text(image, draw_point, text, bg_color, thickness=thickness, drawType=drawType)
     return image
 
@@ -1599,11 +1628,11 @@ def get_font_type(size, font=""):
 
 def cv2_putText(img, text, point, fontFace=None, fontScale=0.8, color=(255, 0, 0), thickness=None):
     # cv2.putText(img, str(text), point, fontFace, fontScale, color=color, thickness=thickness)
-    text_size, baseline = cv2.getTextSize(str(text), cv2.FONT_HERSHEY_SIMPLEX, fontScale, abs(thickness))
+    text_wh, baseline = cv2.getTextSize(str(text), cv2.FONT_HERSHEY_COMPLEX, fontScale, abs(thickness))
     pilimg = Image.fromarray(img)  # Image.fromarray()将数组类型转成图片格式，与np.array()相反
     draw = ImageDraw.Draw(pilimg)  # PIL图片上打印汉字
-    size = text_size[1]  # 字体大小
-    point = (point[0], point[1] - text_size[1] + baseline // 2)
+    size = text_wh[1]  # 字体大小
+    point = (point[0], point[1] - text_wh[1] + baseline // 2)
     font = get_font_type(size=size)
     draw.text(point, text, color, font)
     img[:] = np.asarray(pilimg)
@@ -2333,7 +2362,7 @@ def draw_yaw_pitch_roll_in_right_axis(image,
     cv2.arrowedLine(image, (int(cx), int(cy)), (int(x2), int(y2)), color_pitch_y, 2, tipLength=0.2)
     cv2.arrowedLine(image, (int(cx), int(cy)), (int(x3), int(y3)), color_roll_z, 2, tipLength=0.2)
     if vis:
-        fontFace = cv2.FONT_HERSHEY_SIMPLEX
+        fontFace = cv2.FONT_HERSHEY_COMPLEX
         image = cv2.putText(image, str(text),
                             # (int(cx), int(cy) + 10),
                             (10, 10),
@@ -2379,7 +2408,7 @@ def draw_yaw_pitch_roll_in_left_axis(image,
     cv2.arrowedLine(image, (int(cx), int(cy)), (int(x2), int(y2)), color_pitch_y, 2, tipLength=0.2)
     cv2.arrowedLine(image, (int(cx), int(cy)), (int(x3), int(y3)), color_roll_z, 2, tipLength=0.2)
     if vis:
-        fontFace = cv2.FONT_HERSHEY_SIMPLEX
+        fontFace = cv2.FONT_HERSHEY_COMPLEX
         image = cv2.putText(image, str(text),
                             # (int(cx), int(cy) + 10),
                             (10, 10),
@@ -2701,7 +2730,39 @@ def draw_image_contours(image, contours: List[np.ndarray], texts=[], color=(), a
     """
     colors = [(0, 0, 0)] + colors if colors else color_table
     for i in range(0, len(contours)):
-        c = color if color else colors[(i + 1) % 20]
+        c = color
+        if not color:
+            c = color_label.get_color(texts[i]) if texts else colors[(i + 1) % len(colors)]  # 相同类别相同颜色
+        t = str(texts[i]) if texts else ""
+        p = np.asarray(contours[i], dtype=np.int32)
+        b = (min(p[:, 0]), min(p[:, 1]), max(p[:, 0]), max(p[:, 1]))
+        if len(p.shape) == 2: p = [p]
+        image[:] = cv2.drawContours(image, p, contourIdx=-1, color=c, thickness=thickness)
+        bgimg = image.copy()
+        bgimg = cv2.fillPoly(bgimg, p, color=c)
+        image[:] = cv2.addWeighted(src1=image, alpha=1 - alpha, src2=bgimg, beta=alpha, gamma=0)
+        if t: image[:] = draw_text(image, point=(b[0], b[1]), color=c, text=t, thickness=thickness,
+                                   fontScale=fontScale, drawType=drawType)
+    return image
+
+
+draw_image_contours_texts = draw_image_contours
+
+
+def draw_image_contours_labels_texts(image, contours: List[np.ndarray], labels=[], texts=[], color=(), alpha=0.5,
+                                     thickness=1, fontScale=0.8, drawType="ch", colors=None):
+    """
+    参考：draw_image_mask_color
+    :param image:
+    :param contours: List[np.ndarray],每个列表是一个轮廓(num_points,1,2)
+    :param texts:轮廓文本
+    :param color:绘制轮廓的颜色
+    :param alpha:绘制颜色的透明度 0: 轮廓填充完全透明，1: 轮廓填充完全不透明
+    :param thickness:轮廓线宽
+    :return:
+    """
+    for i in range(0, len(contours)):
+        c = color if color else color_label.get_color(labels[i])  # 相同类别相同颜色
         t = str(texts[i]) if texts else ""
         p = np.asarray(contours[i], dtype=np.int32)
         b = (min(p[:, 0]), min(p[:, 1]), max(p[:, 0]), max(p[:, 1]))