如何在python程序中使用Google Vision API？

在Python程序中使用Google Vision API

基础概念

Google Vision API是Google Cloud提供的一项图像分析服务，能够通过机器学习技术检测图像中的对象、人脸、文本等内容，并提供多种分析功能。

优势

强大的预训练模型：无需自行训练模型即可使用
多种分析功能：包括标签检测、文本识别、人脸检测等
高准确性：基于Google先进的深度学习技术
易于集成：提供简单的REST API和客户端库

准备工作

1. 设置Google Cloud项目

在Google Cloud控制台创建项目
启用Vision API服务
创建服务账号并下载JSON密钥文件

2. 安装客户端库

pip install --upgrade google-cloud-vision

基本使用示例

1. 文本检测

from google.cloud import vision

def detect_text(path):
    """检测图像中的文本"""
    client = vision.ImageAnnotatorClient.from_service_account_file('your-service-account.json')
    
    with open(path, 'rb') as image_file:
        content = image_file.read()

    image = vision.Image(content=content)
    response = client.text_detection(image=image)
    texts = response.text_annotations

    for text in texts:
        print(f'\n"{text.description}"')
        vertices = [(v.x, v.y) for v in text.bounding_poly.vertices]
        print(f'bounds: {vertices}')

    if response.error.message:
        raise Exception(f'{response.error.message}')

detect_text('path/to/your/image.jpg')

2. 标签检测

def detect_labels(path):
    """检测图像中的标签"""
    client = vision.ImageAnnotatorClient.from_service_account_file('your-service-account.json')
    
    with open(path, 'rb') as image_file:
        content = image_file.read()

    image = vision.Image(content=content)
    response = client.label_detection(image=image)
    labels = response.label_annotations

    print('Labels:')
    for label in labels:
        print(f'{label.description} ({label.score*100:.2f}%)')

detect_labels('path/to/your/image.jpg')

3. 人脸检测

def detect_faces(path):
    """检测图像中的人脸"""
    client = vision.ImageAnnotatorClient.from_service_account_file('your-service-account.json')
    
    with open(path, 'rb') as image_file:
        content = image_file.read()

    image = vision.Image(content=content)
    response = client.face_detection(image=image)
    faces = response.face_annotations

    print('Faces:')
    for face in faces:
        print(f'Detection confidence: {face.detection_confidence}')
        print(f'Joy likelihood: {face.joy_likelihood}')
        print(f'Sorrow likelihood: {face.sorrow_likelihood}')
        print(f'Anger likelihood: {face.anger_likelihood}')
        print(f'Surprise likelihood: {face.surprise_likelihood}')

detect_faces('path/to/your/image.jpg')

高级功能

1. 批量处理图像

def batch_annotate_images(input_uris):
    """批量处理图像"""
    client = vision.ImageAnnotatorClient.from_service_account_file('your-service-account.json')
    
    requests = []
    for uri in input_uris:
        request = {
            'image': {'source': {'image_uri': uri}},
            'features': [
                {'type_': vision.Feature.Type.FACE_DETECTION},
                {'type_': vision.Feature.Type.LABEL_DETECTION}
            ]
        }
        requests.append(request)
    
    response = client.batch_annotate_images(requests)
    
    for image_response in response.responses:
        print(f'Labels: {[label.description for label in image_response.label_annotations]}')
        print(f'Faces found: {len(image_response.face_annotations)}')

batch_annotate_images(['gs://your-bucket/image1.jpg', 'gs://your-bucket/image2.jpg'])

2. 使用远程图像

def detect_text_uri(uri):
    """从远程URI检测文本"""
    client = vision.ImageAnnotatorClient.from_service_account_file('your-service-account.json')
    
    image = vision.Image()
    image.source.image_uri = uri

    response = client.text_detection(image=image)
    texts = response.text_annotations

    for text in texts:
        print(f'\n"{text.description}"')

    if response.error.message:
        raise Exception(f'{response.error.message}')

detect_text_uri('https://example.com/image.jpg')