树莓派5B/4B部署yolov11

首先按照yolov8的环境部署

https://www.cnblogs.com/gooutlook/p/18511319

 

教程

https://docs.ultralytics.com/zh/guides/raspberry-pi/#convert-model-to-ncnn-and-run-inference

 

1更新软件包列表，安装 pip 并升级到最新版本
sudo apt update
sudo apt install python3-pip -y
pip install -U pip


2安装 ultralytics pip 软件包与可选依赖项
pip install ultralytics[export]


3重启设备
sudo reboot

　　

在 Raspberry Pi 上使用NCNN

在Ultralytics.NET 支持的所有模型导出格式中，.NET 是最受欢迎的、 NCNN由于NCNN 针对移动/嵌入式平台（如 ARM 架构）进行了高度优化，因此在使用 Raspberry Pi 设备时可提供最佳推理性能。

将模型转换为NCNN 并运行推理

PyTorch 格式的 YOLO11n 模型被转换为NCNN ，以便使用导出的模型进行推理。

 

from ultralytics import YOLO

# Load a YOLO11n PyTorch model
model = YOLO("yolo11n.pt")

# Export the model to NCNN format
model.export(format="ncnn")  # creates 'yolo11n_ncnn_model'

# Load the exported NCNN model
ncnn_model = YOLO("yolo11n_ncnn_model")

# Run inference
results = ncnn_model("https://ultralytics.com/images/bus.jpg")

　　Ultralytics 基准

from ultralytics import YOLO

# Load a YOLO11n PyTorch model
model = YOLO("yolo11n.pt")

# Benchmark YOLO11n speed and accuracy on the COCO8 dataset for all all export formats
results = model.benchmark(data="coco8.yaml", imgsz=640)

　　

下载测试模型

https://docs.ultralytics.com/zh/models/yolo11/#overview

 

 

 https://huggingface.co/Ultralytics/YOLO11/blob/365ed86341e7a7456dbc4cafc09f138814ce9ff1/yolo11n.pt

 

 

多线程测试代码

 

 

 

 

import cv2
import threading
import numpy as np

from ultralytics import YOLO
import datetime

use_ncnn = 1
# 加载YOLOv8模型
model = YOLO('./yolo11n.pt')

# Export the model to NCNN format
model.export(format="ncnn")  # creates 'yolo11n_ncnn_model'

# Load the exported NCNN model
ncnn_model = YOLO("yolo11n_ncnn_model")



# 共享内存类
class SharedMemory:
    def __init__(self):
        self.frame = None
        self.lock = threading.Lock()
        self.running = True  # 添加运行状态标志

# 从 USB 相机读取图像的线程
def capture_thread(shared_memory):
    cap = cv2.VideoCapture(0)
    if not cap.isOpened():
        print("无法打开相机")
        return

    while shared_memory.running:
        ret, frame = cap.read()
        if not ret:
            print("无法读取图像")
            break

        # 将图像存储到共享内存
        with shared_memory.lock:
            shared_memory.frame = frame.copy()

    cap.release()

# 从共享内存读取图像并展示的线程
def display_thread(shared_memory):

    while shared_memory.running:
        with shared_memory.lock:
            if shared_memory.frame is not None:

                global results
                if use_ncnn==1:
                    results = ncnn_model(shared_memory.frame)
                else:
                # 使用YOLOv8进行检测
                    results = model(shared_memory.frame)
                
                
                frame = shared_memory.frame
                # 解析结果并绘制检测框
                for result in results:
                    boxes = result.boxes
                    for box in boxes:
                        x1, y1, x2, y2 = map(int, box.xyxy[0])  # 获取检测框坐标
                        confidence = box.conf[0]  # 获取置信度
                        cls = int(box.cls[0])  # 获取类别
                        label = model.names[cls]  # 获取类别名称

                        # 绘制检测框和标签
                        cv2.rectangle(frame, (x1, y1), (x2, y2), (0, 255, 0), 2)
                        cv2.putText(frame, f"{label} {confidence:.2f}", (x1, y1 - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 2)

                # 获取当前时间
                now = datetime.datetime.now()
                time_str = now.strftime("%Y-%m-%d %H:%M:%S")
                cv2.putText(frame, time_str, (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 255, 0), 2, cv2.LINE_AA)


                cv2.imshow('Camera Feed', frame)
    
        # 检查按键是否为 'q'
        if cv2.waitKey(1) & 0xFF == ord('q'):
            shared_memory.running = False  # 设置运行状态为 False
            break

    cv2.destroyAllWindows()

# 主函数
if __name__ == "__main__":
    shared_memory = SharedMemory()

    # 启动捕获线程
    capture_thread = threading.Thread(target=capture_thread, args=(shared_memory,))
    capture_thread.start()

    # 启动显示线程
    display_thread = threading.Thread(target=display_thread, args=(shared_memory,))
    display_thread.start()

    capture_thread.join()
    display_thread.join()