From a703526ef54ce2fe8c4769c3c245c3da6bce1fcc Mon Sep 17 00:00:00 2001 From: pos97em56 <10225101485@ecnu.stu.edu.cn> Date: Wed, 3 Jul 2024 17:26:27 +0800 Subject: [PATCH] ADD file via upload --- advanced/track/animal_track.py | 160 +++++++++++++++++++++++++++++++++ 1 file changed, 160 insertions(+) create mode 100644 advanced/track/animal_track.py diff --git a/advanced/track/animal_track.py b/advanced/track/animal_track.py new file mode 100644 index 0000000..cd17ee3 --- /dev/null +++ b/advanced/track/animal_track.py @@ -0,0 +1,160 @@ +import tkinter as tk +from tkinter import filedialog +import numpy as np +import cv2 +import os +import time + +# 获取当前脚本文件的目录 +base_path = os.path.dirname(os.path.abspath(__file__)) + +def select_file_and_run(): + """ + 弹出文件选择对话框，选择视频文件并运行检测和追踪程序。 + """ + file_path = filedialog.askopenfilename(filetypes=[("Video files", "*.mp4;*.avi")]) + if file_path: + run_detection_and_tracking(file_path) + +def run_detection_and_tracking(file_path): + """ + 运行目标检测和多目标追踪程序。 + 参数： + file_path - 视频文件路径 + """ + # 加载YOLO模型 + model_path1 = os.path.join(base_path, 'yolo/yolov3.weights') # 权重文件路径 + model_path2 = os.path.join(base_path, 'yolo/yolov3.cfg') # 配置文件路径 + model_path3 = os.path.join(base_path, 'yolo/coco.names') # 类别名称文件路径 + net = cv2.dnn.readNet(model_path1, model_path2) # 加载YOLO模型 + layer_names = net.getLayerNames() # 获取模型层名称 + output_layers = [layer_names[i - 1] for i in net.getUnconnectedOutLayers()] # 获取输出层名称 + classes = [] + with open(model_path3, "r") as f: + classes = [line.strip() for line in f.readlines()] # 读取类别名称 + + # 初始化视频捕捉 + cap = cv2.VideoCapture(file_path) # 打开视频文件 + fps = cap.get(cv2.CAP_PROP_FPS) # 获取视频帧率 + speed_up_factor = 2 # 加速倍数 + delay = int(1000 / fps / speed_up_factor) # 计算加速后每帧之间的时间间隔，以毫秒为单位 + + # 创建多目标追踪器 + trackers = [] + # 创建用于存储已追踪对象信息的列表 + tracked_objects = [] + + while True: + ret, frame = cap.read() # 读取视频帧 + if not ret: + break + + start_time = time.time() + + # 检测目标 + height, width, channels = frame.shape + blob = cv2.dnn.blobFromImage(frame, 0.00392, (416, 416), (0, 0, 0), True, crop=False) + net.setInput(blob) # 设置输入数据 + outs = net.forward(output_layers) # 执行前向传播，获取检测结果 + + # 获取检测结果 + new_objects = [] # 用于存储当前帧检测到的新对象 + + class_ids = [] # 存储检测到的目标的类别ID + confidences = [] # 存储检测到的目标的置信度 + boxes = [] # 存储检测到的目标的边界框 + for out in outs: + for detection in out: + scores = detection[5:] # 获取每个检测框的所有类别得分 + class_id = np.argmax(scores) # 获取得分最高的类别ID，即检测到的物体类别 + confidence = scores[class_id] # 获取该类别的置信度 + if confidence > 0.5: # 筛选置信度超过0.5的检测框 + # 目标检测 + center_x = int(detection[0] * width) # 中心点x坐标 + center_y = int(detection[1] * height) # 中心点y坐标 + w = int(detection[2] * width) # 边界框宽度 + h = int(detection[3] * height) # 边界框高度 + x = int(center_x - w / 2) # 左上角x坐标 + y = int(center_y - h / 2) # 左上角y坐标 + boxes.append([x, y, w, h]) # 存储边界框 + confidences.append(float(confidence)) # 存储置信度 + class_ids.append(class_id) # 存储类别ID + new_objects.append((x, y, w, h, class_id)) # 存储新检测到的对象信息 + + # 非最大值抑制 + indices = cv2.dnn.NMSBoxes(boxes, confidences, 0.5, 0.4) + + # 更新追踪器或添加新的追踪器 + for i in indices.flatten(): # 遍历NMS后的检测框索引 + x, y, w, h = boxes[i] # 获取检测框的左上角坐标和宽高 + label = str(classes[class_ids[i]]) # 获取检测框的类别标签 + if label in ["dog", "cat", "bird", "horse", "sheep", "cow", "elephant", "bear", "zebra", + "giraffe"]: # 只选择动物目标 + # 检查是否已经有相似的追踪器在追踪相同类型的对象 + found_similar = False + for tracked_object in tracked_objects: + if tracked_object[4] == class_ids[i]: # 检查类别是否相同 + # 计算当前检测到的对象与已有追踪器的距离或重叠度 + existing_bbox = (tracked_object[0], tracked_object[1], tracked_object[0] + tracked_object[2], + tracked_object[1] + tracked_object[3]) + new_bbox = (x, y, x + w, y + h) + overlap_area = calculate_overlap(existing_bbox, new_bbox) + if overlap_area > 0.5: # 如果重叠度超过阈值，认为是同一个对象，不再重复追踪 + found_similar = True + break + if not found_similar: + tracker = cv2.TrackerKCF_create() # 创建KCF追踪器 + trackers.append(tracker) # 将追踪器添加到追踪器列表 + trackers[-1].init(frame, (x, y, w, h)) # 初始化追踪器 + tracked_objects.append((x, y, w, h, class_ids[i])) # 添加到已追踪对象列表 + + # 绘制追踪框 + for tracker in trackers: + success, bbox = tracker.update(frame) # 更新追踪器 + if success: + # 画出追踪框 + p1 = (int(bbox[0]), int(bbox[1])) + p2 = (int(bbox[0] + bbox[2]), int(bbox[1] + bbox[3])) + cv2.rectangle(frame, p1, p2, (255, 0, 0), 2) + else: + # 追踪失败 + cv2.putText(frame, "Tracking failure detected", (100, 80), cv2.FONT_HERSHEY_SIMPLEX, + 0.75, (0, 0, 255),2) + + # 显示结果 + cv2.imshow('Tracking', frame) + + # 按下ESC键退出或关闭窗口退出 + if cv2.waitKey(delay) & 0xFF == 27: + break + if cv2.getWindowProperty('Tracking', cv2.WND_PROP_VISIBLE) < 1: + break + + cap.release() + cv2.destroyAllWindows() + +def calculate_overlap(bbox1, bbox2): + """ + 计算两个边界框的重叠面积比（IoU）。 + 参数： + bbox1, bbox2 - 分别是 (x1, y1, x2, y2) 格式的边界框坐标，其中 (x1, y1) 是左上角坐标，(x2, y2) 是右下角坐标 + 返回值： + iou - 重叠面积比 + """ + # 计算交集部分的坐标 + inter_x1 = max(bbox1[0], bbox2[0]) + inter_y1 = max(bbox1[1], bbox2[1]) + inter_x2 = min(bbox1[2], bbox2[2]) + inter_y2 = min(bbox1[3], bbox2[3]) + + # 计算交集区域的面积 + inter_area = max(0, inter_x2 - inter_x1 + 1) * max(0, inter_y2 - inter_y1 + 1) + + # 计算各自的区域面积 + area_bbox1 = (bbox1[2] - bbox1[0] + 1) * (bbox1[3] - bbox1[1] + 1) + area_bbox2 = (bbox2[2] - bbox2[0] + 1) * (bbox2[3] - bbox2[1] + 1) + + # 计算并返回重叠区域的IoU + iou = inter_area / float(area_bbox1 + area_bbox2 - inter_area) + + return iou