Python实现视频中人物和车辆的识别与统计

要实现视频中人物和车辆的识别并统计数量，我们可以使用OpenCV结合深度学习模型（如YOLO或SSD）来实现。下面是一个完整的解决方案：

方案概述

1. 使用OpenCV读取视频帧

2. 加载预训练的目标检测模型（这里使用YOLOv3）

3. 对每帧进行目标检测

4. 过滤出"person"和"car"类别

5. 统计数量并显示结果

完整代码实现

【python】
import cv2
import numpy as np

加载YOLOv3模型

def load_yolo():
net = cv2.dnn.readNet(“yolov3.weights”, “yolov3.cfg”)
classes = []
with open(“coco.names”, “r”) as f:
classes = [line.strip() for line in f.readlines()]

layer_names = net.getLayerNames()
output_layers = [layer_names[i[0] - 1] for i in net.getUnconnectedOutLayers()]
return net, classes, output_layers

检测对象函数

def detect_objects(img, net, output_layers, classes):
height, width, channels = img.shape

# 预处理图像
blob = cv2.dnn.blobFromImage(img, 1/255.0, (416, 416), swapRB=True, crop=False)
net.setInput(blob)
outs = net.forward(output_layers)

# 解析检测结果
class_ids = []
confidences = []
boxes = []

for out in outs:
    for detection in out:
        scores = detection[5:]
        class_id = np.argmax(scores)
        confidence = scores[class_id]
        if confidence > 0.5 and (classes[class_id] == 'person' or classes[class_id] == 'car'):
            # 检测到人或车
            center_x = int(detection[0] * width)
            center_y = int(detection[1] * height)
            w = int(detection[2] * width)
            h = int(detection[3] * height)
            
            # 矩形坐标
            x = int(center_x - w / 2)
            y = int(center_y - h / 2)
            
            boxes.append([x, y, w, h])
            confidences.append(float(confidence))
            class_ids.append(class_id)

# 应用非极大值抑制
indexes = cv2.dnn.NMSBoxes(boxes, confidences, 0.5, 0.4)

return boxes, confidences, class_ids, indexes, classes

主函数

def main():
# 加载模型
model, classes, output_layers = load_yolo()

# 初始化计数器
person_count = 0
car_count = 0
total_frames = 0

# 打开视频文件或摄像头
cap = cv2.VideoCapture("your_video.mp4")  # 替换为你的视频路径或0表示摄像头

if not cap.isOpened():
    print("无法打开视频源")
    return

while True:
    ret, frame = cap.read()
    if not ret:
        break
    
    total_frames += 1
    
    # 检测对象
    boxes, confidences, class_ids, indexes, classes = detect_objects(frame, model, output_layers, classes)
    
    # 临时计数器
    temp_person = 0
    temp_car = 0
    
    # 绘制检测框并计数
    font = cv2.FONT_HERSHEY_PLAIN
    for i in range(len(boxes)):
        if i in indexes:
            x, y, w, h = boxes[i]
            label = str(classes[class_ids[i]])
            
            if label == 'person':
                temp_person += 1
                color = (0, 255, 0)  # 绿色表示人
            elif label == 'car':
                temp_car += 1
                color = (255, 0, 0)  # 蓝色表示车
            
            cv2.rectangle(frame, (x, y), (x + w, y + h), color, 2)
            cv2.putText(frame, label, (x, y + 30), font, 3, color, 3)
    
    # 更新总计数
    person_count += temp_person
    car_count += temp_car
    
    # 显示统计信息
    cv2.putText(frame, f"Total Persons: {person_count}", (10, 30), font, 2, (0, 255, 0), 2)
    cv2.putText(frame, f"Total Cars: {car_count}", (10, 70), font, 2, (255, 0, 0), 2)
    cv2.putText(frame, f"Frame: {total_frames}", (10, 110), font, 2, (255, 255, 255), 2)
    
    # 显示帧
    cv2.imshow("Object Detection", frame)
    
    # 按q退出
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

cap.release()
cv2.destroyAllWindows()

print(f"视频处理完成！总帧数: {total_frames}")
print(f"检测到的人物总数: {person_count}")
print(f"检测到的车辆总数: {car_count}")

if name == “main”:
main()

使用说明

1. 你需要下载以下YOLOv3模型文件：

   • yolov3.weights(权重文件)

   • yolov3.cfg(配置文件)

   • coco.names(包含类别名称的文件)

2. 将这些文件放在与脚本相同的目录下，或者修改代码中的路径。

3. 将your_video.mp4替换为你的视频文件路径，或者使用0表示使用摄像头。

替代方案（使用OpenCV内置模型）

如果你不想使用YOLO，OpenCV也提供了内置的SSD模型：

【python】
def load_opencv_model():
# 加载OpenCV内置的SSD模型
net = cv2.dnn.readNetFromTensorflow(“opencv_extra/testdata/dnn/frozen_inference_graph.pb”,
“opencv_extra/testdata/dnn/ssd_mobilenet_v2_coco_2018_03_29.pbtxt”)
return net

修改后的detect_objects函数

def detect_objects_opencv(img, net):
height, width = img.shape[:2]

# 预处理图像
blob = cv2.dnn.blobFromImage(img, size=(300, 300), swapRB=True, crop=False)
net.setInput(blob)
detections = net.forward()

# 解析检测结果
boxes = []
confidences = []
class_ids = []

for i in range(detections.shape[2]):
    confidence = detections[0, 0, i, 2]
    if confidence > 0.5:
        class_id = int(detections[0, 0, i, 1])
        # 只保留人和车
        if class_id == 1 or class_id == 3:  # 1: person, 3: car
            box = detections[0, 0, i, 3:7] * np.array([width, height, width, height])
            (startX, startY, endX, endY) = box.astype("int")
            boxes.append([startX, startY, endX - startX, endY - startY])
            confidences.append(float(confidence))
            class_ids.append(class_id)

# 应用非极大值抑制
indexes = cv2.dnn.NMSBoxes(boxes,