Raspberry Pi カメラモジュールV3をPython OpenCVで操作する方法

lsb_release -a

sudo python -m pip install --upgrade pip

sudo pip3 install opencv-python==4.5.1.48

sudo pip3 install numpy==1.23.1

sudo apt update

sudo apt install libatlas3-base

libcamera-hello --timeout 0

from picamera2 import Picamera2

picam2 = Picamera2()

# フル解像度で画像を取得するための設定
capture_config = picam2.create_still_configuration(
    main={"size": picam2.camera_properties['PixelArraySize']}  # センサーのフル解像度を指定
)

# 設定の適用
picam2.configure(capture_config)

# カメラの開始
picam2.start()

# 画像のキャプチャ
picam2.capture_file("test.jpg")

# カメラの停止
picam2.stop()

from picamera2 import Picamera2
import datetime

# 現在の日時を取得し、ファイル名に使用する
dt_now = datetime.datetime.now()
file_name = dt_now.strftime('%Y%m%d_%H%M%S')

picam2 = Picamera2()

# フル解像度で画像を取得するための設定
capture_config = picam2.create_still_configuration(
    main={"size": picam2.camera_properties['PixelArraySize']}  # センサーのフル解像度を指定
)

# 設定の適用
picam2.configure(capture_config)

# カメラの開始
picam2.start()

# 画像のキャプチャ
picam2.capture_file(file_name + '.jpg')

# カメラの停止
picam2.stop()

from picamera2 import Picamera2

picam2 = Picamera2()

# 解像度を1920x1080に設定
video_config = picam2.create_video_configuration(
    main={"size": (1920, 1080)}  # フルHDに変更
)

# 設定の適用
picam2.configure(video_config)

# 動画撮影の開始と5秒後の自動停止
picam2.start_and_record_video("test.mp4", duration=5)

# カメラの停止
picam2.stop()

import cv2
from picamera2 import Picamera2
from libcamera import controls

picam2 = Picamera2()
picam2.configure(picam2.create_preview_configuration(main={"format": 'XRGB8888', "size": (640, 480)}))
picam2.start()
#カメラを連続オートフォーカスモードにする
picam2.set_controls({"AfMode": controls.AfModeEnum.Continuous})

while True:
  im = picam2.capture_array()
  cv2.imshow("Camera", im)
 
  key = cv2.waitKey(1)
  # Escキーを入力されたら画面を閉じる
  if key == 27:
    break

picam2.stop()
cv2.destroyAllWindows()

picam2.set_controls({"AfMode": controls.AfModeEnum.Continuous, "AfSpeed": controls.AfSpeedEnum.Fast})

import cv2
from picamera2 import Picamera2
from libcamera import controls
import os

picam2 = Picamera2()

#HDRをオンにする
os.system("v4l2-ctl --set-ctrl wide_dynamic_range=1 -d /dev/v4l-subdev0")
print("Setting HDR to ON")

picam2.configure(picam2.create_preview_configuration(main={"format": 'XRGB8888', "size": (640, 480)}))
picam2.start()
#カメラを連続オートフォーカスモードにする
picam2.set_controls({"AfMode": controls.AfModeEnum.Continuous, "AfSpeed": controls.AfSpeedEnum.Fast})

while True:
  im = picam2.capture_array()
  cv2.imshow("Camera", im)
 
  key = cv2.waitKey(1)
  # Escキーを入力されたら画面を閉じる
  if key == 27:
    break

picam2.stop()
cv2.destroyAllWindows()

#HDRをオフにする
os.system("v4l2-ctl --set-ctrl wide_dynamic_range=0 -d /dev/v4l-subdev0")
print("Setting HDR to OFF")

import cv2
from picamera2 import Picamera2
from libcamera import controls

face_detector = cv2.CascadeClassifier("/usr/local/lib/python3.9/dist-packages/cv2/data/"\
"haarcascade_frontalface_default.xml")
cv2.startWindowThread()

picam2 = Picamera2()
picam2.configure(picam2.create_preview_configuration(main={"format": 'XRGB8888', "size": (640, 480)}))
picam2.start()
#カメラを連続オートフォーカスモードにする
picam2.set_controls({"AfMode": controls.AfModeEnum.Continuous})

while True:
    im = picam2.capture_array()

    grey = cv2.cvtColor(im, cv2.COLOR_BGR2GRAY)
    faces = face_detector.detectMultiScale(grey, 1.1, 5)

    for (x, y, w, h) in faces:
        cv2.rectangle(im, (x, y), (x + w, y + h), (0, 255, 0))

    cv2.imshow("Camera", im)
   
    key = cv2.waitKey(1)
    # Escキーを入力されたら画面を閉じる
    if key == 27:
        break

picam2.stop()
cv2.destroyAllWindows()

sudo apt install libcap-dev

git clone https://github.com/googlesamples/mediapipe.git

cd mediapipe/examples/object_detection/raspberry_pi

sh setup.sh

# Copyright 2023 The MediaPipe Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import argparse
import sys
import time
import cv2
import mediapipe as mp
from picamera2 import Picamera2
from libcamera import controls
from mediapipe.tasks import python
from mediapipe.tasks.python import vision
from utils import visualize

# Initial configuration of the camera
picam2 = Picamera2()
picam2.configure(picam2.create_preview_configuration(
    main={"format": 'XRGB8888', "size": (640, 480)}))
picam2.start()
picam2.set_controls({"AfMode": controls.AfModeEnum.Continuous})

# Global variables to calculate FPS
COUNTER, FPS = 0, 0
START_TIME = time.time()

def run(model: str, max_results: int, score_threshold: float,
        width: int, height: int) -> None:
            
    global is_inference_in_flight
    global latest_detection_result

    # Initialize global variables here
    is_inference_in_flight = False
    latest_detection_result = None            

    def save_result(result: vision.ObjectDetectorResult,
                    unused_output_image: mp.Image,
                    timestamp_ms: int):
        """Callback when the detection is done."""
        global is_inference_in_flight, latest_detection_result, COUNTER, START_TIME, FPS

        # Mark that inference has finished, so we can send another frame
        is_inference_in_flight = False

        # Update FPS
        if COUNTER % fps_avg_frame_count == 0:
            current_time = time.time()
            FPS = fps_avg_frame_count / (current_time - START_TIME)
            START_TIME = current_time

        latest_detection_result = result
        COUNTER += 1

    # Initialize the object detection model
    base_options = python.BaseOptions(model_asset_path=model)
    options = vision.ObjectDetectorOptions(
        base_options=base_options,
        running_mode=vision.RunningMode.LIVE_STREAM,
        max_results=max_results,
        score_threshold=score_threshold,
        result_callback=save_result
    )
    detector = vision.ObjectDetector.create_from_options(options)

    # Visualization parameters
    row_size = 50  # pixels
    left_margin = 24  # pixels
    text_color = (0, 0, 0)  # black
    font_size = 1
    font_thickness = 1
    fps_avg_frame_count = 10

    # Create a window to display the detection results
    cv2.namedWindow('object_detection', cv2.WINDOW_NORMAL)
    cv2.resizeWindow('object_detection', 800, 600)

    while True:
        frame = picam2.capture_array()
        if frame is None:
            break

        # Resize and convert for inference
        image = cv2.resize(frame, (width, height))
        rgb_image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
        mp_image = mp.Image(image_format=mp.ImageFormat.SRGB, data=rgb_image)

        # If the previous inference is done, start a new one
        if not is_inference_in_flight:
            detector.detect_async(mp_image, time.time_ns() // 1_000_000)
            is_inference_in_flight = True
        else:
            # If inference is still in flight,
            # we do NOT enqueue a new frame to avoid backlog
            pass

        # Draw FPS
        fps_text = 'FPS = {:.1f}'.format(FPS)
        text_location = (left_margin, row_size)
        cv2.putText(image, fps_text, text_location, cv2.FONT_HERSHEY_DUPLEX,
                    font_size, text_color, font_thickness, cv2.LINE_AA)

        # Visualize the latest detection result (if it exists)
        if latest_detection_result is not None:
            image = visualize(image, latest_detection_result)

        cv2.imshow('object_detection', image)

        # Stop if ESC is pressed
        if cv2.waitKey(1) == 27:
            break

    detector.close()
    picam2.stop()
    cv2.destroyAllWindows()

def main():
    parser = argparse.ArgumentParser(
        formatter_class=argparse.ArgumentDefaultsHelpFormatter)
    parser.add_argument('--model', help='Path of the object detection model.',
                        required=False, default='efficientdet.tflite')
    parser.add_argument('--maxResults', help='Max number of detection results.',
                        required=False, default=5)
    parser.add_argument('--scoreThreshold',
                        help='The score threshold of detection results.',
                        required=False, type=float, default=0.25)
    parser.add_argument('--cameraId', help='Id of camera.',
                        required=False, type=int, default=0)
    parser.add_argument('--frameWidth',
                        help='Width of frame to capture from camera.',
                        required=False, type=int, default=1280)
    parser.add_argument('--frameHeight',
                        help='Height of frame to capture from camera.',
                        required=False, type=int, default=720)
    args = parser.parse_args()

    run(args.model, int(args.maxResults), args.scoreThreshold,
        args.frameWidth, args.frameHeight)

if __name__ == '__main__':
    main()

# Copyright 2023 The MediaPipe Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import argparse
import sys
import time
import cv2
import mediapipe as mp
from picamera2 import Picamera2
from libcamera import controls
from mediapipe.tasks import python
from mediapipe.tasks.python import vision
from utils import visualize

# 任意の物体名を指定する変数（ここで変更可能）
target_object = "person"  # ここを好きな物体名に変更できる

# カメラの初期設定
picam2 = Picamera2()
picam2.configure(picam2.create_preview_configuration(
    main={"format": 'XRGB8888', "size": (640, 480)}))
picam2.start()
picam2.set_controls({"AfMode": controls.AfModeEnum.Continuous})

# FPS計算用のグローバル変数
COUNTER, FPS = 0, 0
START_TIME = time.time()
is_inference_in_flight = False
latest_detection_result = None
fps_avg_frame_count = 10

def save_result(result: vision.ObjectDetectorResult,
                unused_output_image: mp.Image,
                timestamp_ms: int):
    """検出が完了したときに呼び出されるコールバック関数"""
    global is_inference_in_flight, latest_detection_result, COUNTER, START_TIME, FPS

    # 推論が終了したことをマークし、新しいフレームを送信できるようにする
    is_inference_in_flight = False

    # 検出結果から指定した物体の数をカウントする
    object_count = sum(1 for detection in result.detections if detection.categories[0].category_name == target_object)
    
    # 検出された物体の数をコンソールに出力
    print(f"検出された {target_object} の数: {object_count}")

    # FPSを更新
    if COUNTER % fps_avg_frame_count == 0:
        current_time = time.time()
        FPS = fps_avg_frame_count / (current_time - START_TIME)
        START_TIME = current_time

    latest_detection_result = result
    COUNTER += 1

    # 画像上に検出された物体の数を描画
    object_text = f'{target_object.capitalize()}: {object_count}'
    image_copy = unused_output_image.numpy_view().copy()  # 書き込み可能なコピーを作成
    cv2.putText(image_copy, object_text, (24, 100),
                cv2.FONT_HERSHEY_DUPLEX, 1, (0, 0, 255), 1, cv2.LINE_AA)

def run(model: str, max_results: int, score_threshold: float,
        width: int, height: int) -> None:
    global is_inference_in_flight

    # オブジェクト検出モデルを初期化
    base_options = python.BaseOptions(model_asset_path=model)
    options = vision.ObjectDetectorOptions(
        base_options=base_options,
        running_mode=vision.RunningMode.LIVE_STREAM,
        max_results=max_results,
        score_threshold=score_threshold,
        result_callback=save_result
    )
    detector = vision.ObjectDetector.create_from_options(options)

    # 表示に関するパラメータ
    row_size = 50  # ピクセル
    left_margin = 24  # ピクセル
    text_color = (0, 0, 0)  # 黒
    font_size = 1
    font_thickness = 1

    # 検出結果を表示するウィンドウを作成
    cv2.namedWindow('object_detection', cv2.WINDOW_NORMAL)
    cv2.resizeWindow('object_detection', 800, 600)

    while True:
        frame = picam2.capture_array()
        if frame is None:
            break

        # フレームをリサイズし、推論用に変換
        image = cv2.resize(frame, (width, height))
        rgb_image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
        mp_image = mp.Image(image_format=mp.ImageFormat.SRGB, data=rgb_image)

        # 前回の推論が終了していれば、新しい推論を開始
        if not is_inference_in_flight:
            detector.detect_async(mp_image, time.time_ns() // 1_000_000)
            is_inference_in_flight = True

        # FPSを画像上に描画
        fps_text = 'FPS = {:.1f}'.format(FPS)
        text_location = (left_margin, row_size)
        cv2.putText(image, fps_text, text_location, cv2.FONT_HERSHEY_DUPLEX,
                    font_size, text_color, font_thickness, cv2.LINE_AA)

        # 最新の検出結果が存在する場合は画像に描画
        if latest_detection_result is not None:
            image = visualize(image, latest_detection_result)

            # 画像上に検出された物体の数を描画
            object_count = sum(1 for detection in latest_detection_result.detections if detection.categories[0].category_name == target_object)
            object_text = f'{target_object.capitalize()}: {object_count}'
            cv2.putText(image, object_text, (24, 100), cv2.FONT_HERSHEY_DUPLEX,
                        font_size, text_color, font_thickness, cv2.LINE_AA)

        cv2.imshow('object_detection', image)

        # ESCキーが押されたら終了
        if cv2.waitKey(1) == 27:
            break

    detector.close()
    picam2.stop()
    cv2.destroyAllWindows()

def main():
    parser = argparse.ArgumentParser(
        formatter_class=argparse.ArgumentDefaultsHelpFormatter)
    parser.add_argument('--model', help='オブジェクト検出モデルのパス',
                        required=False, default='efficientdet.tflite')
    parser.add_argument('--maxResults', help='検出結果の最大数',
                        required=False, default=5)
    parser.add_argument('--scoreThreshold',
                        help='検出結果のスコア閾値',
                        required=False, type=float, default=0.25)
    parser.add_argument('--cameraId', help='カメラID',
                        required=False, type=int, default=0)
    parser.add_argument('--frameWidth',
                        help='カメラからキャプチャするフレームの幅',
                        required=False, type=int, default=1280)
    parser.add_argument('--frameHeight',
                        help='カメラからキャプチャするフレームの高さ',
                        required=False, type=int, default=720)
    args = parser.parse_args()

    run(args.model, int(args.maxResults), args.scoreThreshold,
        args.frameWidth, args.frameHeight)

if __name__ == '__main__':
    main()

sudo apt update

sudo apt install libcap-dev

python3 -m venv my_env --system-site-packages

source my_env/bin/activate

pip install --upgrade pip

pip install opencv-python==4.10.0.84