# Gradio YOLOv8 Det v1.3.0 # 创建人:曾逸夫 # 创建时间:2023-12-15 # pip install gradio>=4.9.1 # python gradio_yolov8_det_v1.py import argparse import csv import random import sys from collections import Counter from pathlib import Path import cv2 import gradio as gr from gradio_imageslider import ImageSlider import tempfile import uuid import numpy as np from matplotlib import font_manager from ultralytics import YOLO ROOT_PATH = sys.path[0] # 项目根目录 # --------------------- 字体库 --------------------- SimSun_path = f"{ROOT_PATH}/fonts/SimSun.ttf" # 宋体文件路径 TimesNesRoman_path = f"{ROOT_PATH}/fonts/TimesNewRoman.ttf" # 新罗马字体文件路径 # 宋体 SimSun = font_manager.FontProperties(fname=SimSun_path, size=12) # 新罗马字体 TimesNesRoman = font_manager.FontProperties(fname=TimesNesRoman_path, size=12) import yaml from PIL import Image, ImageDraw, ImageFont from util.fonts_opt import is_fonts # Gradio YOLOv8 Det版本 GYD_VERSION = "Gradio YOLOv8 Det v1.2.1" # 文件后缀 suffix_list = [".csv", ".yaml"] # 字体大小 FONTSIZE = 25 # 目标尺寸 obj_style = ["小目标", "中目标", "大目标"] GYD_TITLE = """

Simple Icons

基于 Gradio 的 YOLOv8 通用计算机视觉演示系统

集成目标检测、图像分割和图像分类于一体,可自定义检测模型

Fork me on Gitee

""" GYD_SUB_TITLE = """ 作者:曾逸夫,Gitee:https://gitee.com/PyCVer ,Github:https://github.com/Zengyf-CVer """ EXAMPLES_DET = [ ["./img_examples/bus.jpg", "yolov8s", "cpu", 640, 0.6, 0.5, 100, "所有尺寸"], ["./img_examples/giraffe.jpg", "yolov8l", "cpu", 320, 0.5, 0.45, 100, "所有尺寸"], ["./img_examples/zidane.jpg", "yolov8m", "cpu", 640, 0.6, 0.5, 100, "所有尺寸"], [ "./img_examples/Millenial-at-work.jpg", "yolov8x", "cpu", 1280, 0.5, 0.5, 100, "所有尺寸", ], ["./img_examples/bus.jpg", "yolov8s-seg", "cpu", 640, 0.6, 0.5, 100, "所有尺寸"], [ "./img_examples/Millenial-at-work.jpg", "yolov8x-seg", "cpu", 1280, 0.5, 0.5, 100, "所有尺寸", ], ] EXAMPLES_CLAS = [ ["./img_examples/img_clas/ILSVRC2012_val_00000008.JPEG", "yolov8s-cls"], ["./img_examples/img_clas/ILSVRC2012_val_00000018.JPEG", "yolov8l-cls"], ["./img_examples/img_clas/ILSVRC2012_val_00000023.JPEG", "yolov8m-cls"], ["./img_examples/img_clas/ILSVRC2012_val_00000067.JPEG", "yolov8m-cls"], ["./img_examples/img_clas/ILSVRC2012_val_00000077.JPEG", "yolov8m-cls"], ["./img_examples/img_clas/ILSVRC2012_val_00000247.JPEG", "yolov8m-cls"], ] GYD_CSS = """#disp_image { text-align: center; /* Horizontally center the content */ }""" def parse_args(known=False): parser = argparse.ArgumentParser(description=GYD_VERSION) parser.add_argument( "--model_name", "-mn", default="yolov8s", type=str, help="model name" ) parser.add_argument( "--model_cfg", "-mc", default="./model_config/model_name_all.yaml", type=str, help="model config", ) parser.add_argument( "--cls_name", "-cls", default="./cls_name/cls_name_zh.yaml", type=str, help="cls name", ) parser.add_argument( "--cls_imgnet_name", "-cin", default="./cls_name/cls_imagenet_name_zh.yaml", type=str, help="cls ImageNet name", ) parser.add_argument( "--nms_conf", "-conf", default=0.5, type=float, help="model NMS confidence threshold", ) parser.add_argument( "--nms_iou", "-iou", default=0.45, type=float, help="model NMS IoU threshold" ) parser.add_argument( "--inference_size", "-isz", default=640, type=int, help="model inference size" ) parser.add_argument( "--max_detnum", "-mdn", default=50, type=float, help="model max det num" ) parser.add_argument( "--slider_step", "-ss", default=0.05, type=float, help="slider step" ) parser.add_argument( "--is_login", "-isl", action="store_true", default=False, help="is login", ) parser.add_argument( "--usr_pwd", "-up", nargs="+", type=str, default=["admin", "admin"], help="user & password for login", ) parser.add_argument( "--is_share", "-is", action="store_true", default=False, help="is login", ) parser.add_argument( "--server_port", "-sp", default=7860, type=int, help="server port" ) args = parser.parse_known_args()[0] if known else parser.parse_args() return args # yaml文件解析 def yaml_parse(file_path): return yaml.safe_load(open(file_path, encoding="utf-8").read()) # yaml csv 文件解析 def yaml_csv(file_path, file_tag): file_suffix = Path(file_path).suffix if file_suffix == suffix_list[0]: # 模型名称 file_names = [i[0] for i in list(csv.reader(open(file_path)))] # csv版 elif file_suffix == suffix_list[1]: # 模型名称 file_names = yaml_parse(file_path).get(file_tag) # yaml版 else: print(f"{file_path}格式不正确!程序退出!") sys.exit() return file_names # 检查网络连接 def check_online(): # 参考:https://github.com/ultralytics/yolov5/blob/master/utils/general.py # Check internet connectivity import socket try: socket.create_connection(("1.1.1.1", 443), 5) # check host accessibility return True except OSError: return False # 标签和边界框颜色设置 def color_set(cls_num): color_list = [] for i in range(cls_num): color = tuple(np.random.choice(range(256), size=3)) color_list.append(color) return color_list # 随机生成浅色系或者深色系 def random_color(cls_num, is_light=True): color_list = [] for i in range(cls_num): color = ( random.randint(0, 127) + int(is_light) * 128, random.randint(0, 127) + int(is_light) * 128, random.randint(0, 127) + int(is_light) * 128, ) color_list.append(color) return color_list # 检测绘制 def pil_draw(img, score_l, bbox_l, cls_l, cls_index_l, textFont, color_list): img_pil = ImageDraw.Draw(img) id = 0 for score, (xmin, ymin, xmax, ymax), label, cls_index in zip( score_l, bbox_l, cls_l, cls_index_l ): img_pil.rectangle( [xmin, ymin, xmax, ymax], fill=None, outline=color_list[cls_index], width=2 ) # 边界框 countdown_msg = f"{id}-{label} {score:.2f}" # text_w, text_h = textFont.getsize(countdown_msg) # 标签尺寸 pillow 9.5.0 # left, top, left + width, top + height # 标签尺寸 pillow 10.0.0 text_xmin, text_ymin, text_xmax, text_ymax = textFont.getbbox(countdown_msg) # 标签背景 img_pil.rectangle( # (xmin, ymin, xmin + text_w, ymin + text_h), # pillow 9.5.0 ( xmin, ymin, xmin + text_xmax - text_xmin, ymin + text_ymax - text_ymin, ), # pillow 10.0.0 fill=color_list[cls_index], outline=color_list[cls_index], ) # 标签 img_pil.multiline_text( (xmin, ymin), countdown_msg, fill=(0, 0, 0), font=textFont, align="center", ) id += 1 return img # 绘制多边形 def polygon_drawing(img_mask, canvas, color_seg): # ------- RGB转BGR ------- color_seg = list(color_seg) color_seg[0], color_seg[2] = color_seg[2], color_seg[0] color_seg = tuple(color_seg) # 定义多边形的顶点 pts = np.array(img_mask, dtype=np.int32) # 多边形绘制 cv2.drawContours(canvas, [pts], -1, color_seg, thickness=-1) # 输出分割结果 def seg_output(img_path, seg_mask_list, color_list, cls_list): img = cv2.imread(img_path) img_c = img.copy() # w, h = img.shape[1], img.shape[0] # 获取分割坐标 for seg_mask, cls_index in zip(seg_mask_list, cls_list): img_mask = [] for i in range(len(seg_mask)): # img_mask.append([seg_mask[i][0] * w, seg_mask[i][1] * h]) img_mask.append([seg_mask[i][0], seg_mask[i][1]]) polygon_drawing(img_mask, img_c, color_list[int(cls_index)]) # 绘制分割图形 img_mask_merge = cv2.addWeighted(img, 0.3, img_c, 0.7, 0) # 合并图像 return img_mask_merge # 目标检测和图像分割模型加载 def model_det_loading( img_path, device_opt, conf, iou, infer_size, max_det, yolo_model="yolov8n.pt" ): model = YOLO(yolo_model) results = model( source=img_path, device=device_opt, imgsz=infer_size, conf=conf, iou=iou, max_det=max_det, ) results = list(results)[0] return results # 图像分类模型加载 def model_cls_loading(img_path, yolo_model="yolov8s-cls.pt"): model = YOLO(yolo_model) results = model(source=img_path) results = list(results)[0] return results # YOLOv8图片检测函数 def yolo_det_img( img_path, model_name, device_opt, infer_size, conf, iou, max_det, obj_size ): global model, model_name_tmp, device_tmp s_obj, m_obj, l_obj = 0, 0, 0 area_obj_all = [] # 目标面积 score_det_stat = [] # 置信度统计 bbox_det_stat = [] # 边界框统计 cls_det_stat = [] # 类别数量统计 cls_index_det_stat = [] # 1 # 模型加载 predict_results = model_det_loading( img_path, device_opt, conf, iou, infer_size, max_det, yolo_model=f"{model_name}.pt", ) # 检测参数 xyxy_list = predict_results.boxes.xyxy.cpu().numpy().tolist() conf_list = predict_results.boxes.conf.cpu().numpy().tolist() cls_list = predict_results.boxes.cls.cpu().numpy().tolist() # 颜色列表 color_list = random_color(len(model_cls_name_cp), True) img = Image.open(img_path) img_cp = img.copy() # 图像分割 if model_name[-3:] == "seg": # masks_list = predict_results.masks.xyn masks_list = predict_results.masks.xy img_mask_merge = seg_output(img_path, masks_list, color_list, cls_list) img = Image.fromarray(cv2.cvtColor(img_mask_merge, cv2.COLOR_BGRA2RGB)) # 判断检测对象是否为空 if xyxy_list != []: # ---------------- 加载字体 ---------------- yaml_index = cls_name.index(".yaml") cls_name_lang = cls_name[yaml_index - 2 : yaml_index] if cls_name_lang == "zh": # 中文 textFont = ImageFont.truetype( str(f"{ROOT_PATH}/fonts/SimSun.ttf"), size=FONTSIZE ) elif cls_name_lang in ["en", "ru", "es", "ar"]: # 英文、俄语、西班牙语、阿拉伯语 textFont = ImageFont.truetype( str(f"{ROOT_PATH}/fonts/TimesNewRoman.ttf"), size=FONTSIZE ) elif cls_name_lang == "ko": # 韩语 textFont = ImageFont.truetype( str(f"{ROOT_PATH}/fonts/malgun.ttf"), size=FONTSIZE ) for i in range(len(xyxy_list)): # ------------ 边框坐标 ------------ x0 = int(xyxy_list[i][0]) y0 = int(xyxy_list[i][1]) x1 = int(xyxy_list[i][2]) y1 = int(xyxy_list[i][3]) # ---------- 加入目标尺寸 ---------- w_obj = x1 - x0 h_obj = y1 - y0 area_obj = w_obj * h_obj # 目标尺寸 if obj_size == obj_style[0] and area_obj > 0 and area_obj <= 32**2: obj_cls_index = int(cls_list[i]) # 类别索引 cls_index_det_stat.append(obj_cls_index) obj_cls = model_cls_name_cp[obj_cls_index] # 类别 cls_det_stat.append(obj_cls) bbox_det_stat.append((x0, y0, x1, y1)) conf = float(conf_list[i]) # 置信度 score_det_stat.append(conf) area_obj_all.append(area_obj) elif ( obj_size == obj_style[1] and area_obj > 32**2 and area_obj <= 96**2 ): obj_cls_index = int(cls_list[i]) # 类别索引 cls_index_det_stat.append(obj_cls_index) obj_cls = model_cls_name_cp[obj_cls_index] # 类别 cls_det_stat.append(obj_cls) bbox_det_stat.append((x0, y0, x1, y1)) conf = float(conf_list[i]) # 置信度 score_det_stat.append(conf) area_obj_all.append(area_obj) elif obj_size == obj_style[2] and area_obj > 96**2: obj_cls_index = int(cls_list[i]) # 类别索引 cls_index_det_stat.append(obj_cls_index) obj_cls = model_cls_name_cp[obj_cls_index] # 类别 cls_det_stat.append(obj_cls) bbox_det_stat.append((x0, y0, x1, y1)) conf = float(conf_list[i]) # 置信度 score_det_stat.append(conf) area_obj_all.append(area_obj) elif obj_size == "所有尺寸": obj_cls_index = int(cls_list[i]) # 类别索引 cls_index_det_stat.append(obj_cls_index) obj_cls = model_cls_name_cp[obj_cls_index] # 类别 cls_det_stat.append(obj_cls) bbox_det_stat.append((x0, y0, x1, y1)) conf = float(conf_list[i]) # 置信度 score_det_stat.append(conf) area_obj_all.append(area_obj) det_img = pil_draw( img, score_det_stat, bbox_det_stat, cls_det_stat, cls_index_det_stat, textFont, color_list, ) # -------------- 目标尺寸计算 -------------- for i in range(len(area_obj_all)): if 0 < area_obj_all[i] <= 32**2: s_obj = s_obj + 1 elif 32**2 < area_obj_all[i] <= 96**2: m_obj = m_obj + 1 elif area_obj_all[i] > 96**2: l_obj = l_obj + 1 sml_obj_total = s_obj + m_obj + l_obj objSize_dict = {} objSize_dict = { obj_style[i]: [s_obj, m_obj, l_obj][i] / sml_obj_total for i in range(3) } # ------------ 类别统计 ------------ clsRatio_dict = {} clsDet_dict = Counter(cls_det_stat) clsDet_dict_sum = sum(clsDet_dict.values()) for k, v in clsDet_dict.items(): clsRatio_dict[k] = v / clsDet_dict_sum images = (det_img, img_cp) images_names = ("det", "raw") images_path = tempfile.mkdtemp() images_paths = [] uuid_name = uuid.uuid4() for image, image_name in zip(images, images_names): image.save(images_path + f"/img_{uuid_name}_{image_name}.jpg") images_paths.append(images_path + f"/img_{uuid_name}_{image_name}.jpg") gr.Info("图片检测成功!") return (det_img, img_cp), images_paths, objSize_dict, clsRatio_dict else: raise gr.Error("图片检测失败!") # YOLOv8图片分类函数 def yolo_cls_img(img_path, model_name): # 模型加载 predict_results = model_cls_loading(img_path, yolo_model=f"{model_name}.pt") det_img = Image.open(img_path) clas_ratio_list = predict_results.probs.top5conf.tolist() clas_index_list = predict_results.probs.top5 clas_name_list = [] for i in clas_index_list: # clas_name_list.append(predict_results.names[i]) clas_name_list.append(model_cls_imagenet_name_cp[i]) clsRatio_dict = {} index_cls = 0 clsDet_dict = Counter(clas_name_list) for k, v in clsDet_dict.items(): clsRatio_dict[k] = clas_ratio_list[index_cls] index_cls += 1 return det_img, clsRatio_dict def main(args): gr.close_all() global model_cls_name_cp, model_cls_imagenet_name_cp, cls_name nms_conf = args.nms_conf nms_iou = args.nms_iou model_name = args.model_name model_cfg = args.model_cfg cls_name = args.cls_name cls_imagenet_name = args.cls_imgnet_name # ImageNet类别 inference_size = args.inference_size max_detnum = args.max_detnum slider_step = args.slider_step is_fonts(f"{ROOT_PATH}/fonts") # 检查字体文件 model_names = yaml_csv(model_cfg, "model_names") # 模型名称 model_cls_name = yaml_csv(cls_name, "model_cls_name") # 类别名称 model_cls_imagenet_name = yaml_csv(cls_imagenet_name, "model_cls_name") # 类别名称 model_cls_name_cp = model_cls_name.copy() # 类别名称 model_cls_imagenet_name_cp = model_cls_imagenet_name.copy() # 类别名称 custom_theme = gr.themes.Soft(primary_hue="blue").set( button_secondary_background_fill="*neutral_100", button_secondary_background_fill_hover="*neutral_200", ) custom_css = GYD_CSS # ------------ Gradio Blocks ------------ with gr.Blocks(theme=custom_theme, css=custom_css) as gyd: with gr.Row(): gr.Markdown(GYD_TITLE) with gr.Row(): gr.Markdown(GYD_SUB_TITLE) with gr.Row(): with gr.Tabs(): with gr.TabItem("目标检测与图像分割"): with gr.Row(): with gr.Column(scale=1): with gr.Row(): inputs_img = gr.Image(image_mode="RGB", type="filepath", label="原始图片") with gr.Row(): device_opt = gr.Radio(choices=["cpu", "0", "1", "2", "3"], value="cpu", label="设备") with gr.Row(): inputs_model = gr.Dropdown(choices=model_names, value=model_name, type="value", label="模型") with gr.Accordion("高级设置", open=True): with gr.Row(): inputs_size = gr.Slider(320, 1600, step=1, value=inference_size, label="推理尺寸") max_det = gr.Slider(1, 1000, step=1, value=max_detnum, label="最大检测数") with gr.Row(): input_conf = gr.Slider(0, 1, step=slider_step, value=nms_conf, label="置信度阈值") inputs_iou = gr.Slider(0, 1, step=slider_step, value=nms_iou, label="IoU 阈值") with gr.Row(): obj_size = gr.Radio(choices=["所有尺寸", "小目标", "中目标", "大目标"], value="所有尺寸", label="目标尺寸") with gr.Row(): gr.ClearButton(inputs_img, value="清除") det_btn_img = gr.Button(value='检测', variant="primary") with gr.Column(scale=1): # with gr.Row(): # outputs_img = gr.Image(type="pil", label="检测图片") with gr.Row(): outputs_img_slider = ImageSlider(position=0.5, label="检测图片") with gr.Row(): outputs_imgfiles = gr.Files(label="图片下载") with gr.Row(): outputs_objSize = gr.Label(label="目标尺寸占比统计") with gr.Row(): outputs_clsSize = gr.Label(label="类别检测占比统计") with gr.Row(): gr.Examples( examples=EXAMPLES_DET, fn=yolo_det_img, inputs=[ inputs_img, inputs_model, device_opt, inputs_size, input_conf, inputs_iou, max_det, obj_size], # outputs=[outputs_img, outputs_objSize, outputs_clsSize], cache_examples=False) with gr.TabItem("图像分类"): with gr.Row(): with gr.Column(scale=1): with gr.Row(): inputs_img_cls = gr.Image(image_mode="RGB", type="filepath", label="原始图片") with gr.Row(): inputs_model_cls = gr.Dropdown(choices=[ "yolov8n-cls", "yolov8s-cls", "yolov8l-cls", "yolov8m-cls", "yolov8x-cls"], value="yolov8s-cls", type="value", label="模型") with gr.Row(): gr.ClearButton(inputs_img, value="清除") det_btn_img_cls = gr.Button(value='检测', variant="primary") with gr.Column(scale=1): with gr.Row(): outputs_img_cls = gr.Image(type="pil", label="检测图片") with gr.Row(): outputs_ratio_cls = gr.Label(label="图像分类结果") with gr.Row(): gr.Examples( examples=EXAMPLES_CLAS, fn=yolo_cls_img, inputs=[inputs_img_cls, inputs_model_cls], # outputs=[outputs_img_cls, outputs_ratio_cls], cache_examples=False) with gr.Accordion("Gradio YOLOv8 Det 安装与使用教程"): gr.Markdown( """## Gradio YOLOv8 Det 安装与使用教程 ```shell conda create -n yolo python==3.8 conda activate yolo # 进入环境 git clone https://gitee.com/CV_Lab/gradio-yolov8-det.git cd gradio-yolov8-det pip install -r ./requirements.txt -U ``` ```shell # 共享模式 python gradio_yolov8_det_v1.py -is # 在浏览器中以共享模式打开,https://**.gradio.app/ # 自定义模型配置 python gradio_yolov8_det_v1.py -mc ./model_config/model_name_all.yaml # 自定义下拉框默认模型名称 python gradio_yolov8_det_v1.py -mn yolov8m # 自定义类别名称 python gradio_yolov8_det_v1.py -cls ./cls_name/cls_name_zh.yaml (目标检测与图像分割) python gradio_yolov8_det_v1.py -cin ./cls_name/cls_imgnet_name_zh.yaml (图像分类) # 自定义NMS置信度阈值 python gradio_yolov8_det_v1.py -conf 0.8 # 自定义NMS IoU阈值 python gradio_yolov8_det_v1.py -iou 0.5 # 设置推理尺寸,默认为640 python gradio_yolov8_det_v1.py -isz 320 # 设置最大检测数,默认为50 python gradio_yolov8_det_v1.py -mdn 100 # 设置滑块步长,默认为0.05 python gradio_yolov8_det_v1.py -ss 0.01 ``` """ ) det_btn_img.click( fn=yolo_det_img, inputs=[ inputs_img, inputs_model, device_opt, inputs_size, input_conf, inputs_iou, max_det, obj_size, ], outputs=[ outputs_img_slider, outputs_imgfiles, outputs_objSize, outputs_clsSize, ], ) det_btn_img_cls.click( fn=yolo_cls_img, inputs=[inputs_img_cls, inputs_model_cls], outputs=[outputs_img_cls, outputs_ratio_cls], ) return gyd if __name__ == "__main__": args = parse_args() gyd = main(args) is_share = args.is_share gyd.queue().launch( inbrowser=True, # 自动打开默认浏览器 share=is_share, # 项目共享,其他设备可以访问 favicon_path="./icon/logo.ico", # 网页图标 show_error=True, # 在浏览器控制台中显示错误信息 quiet=True, # 禁止大多数打印语句 )