21、目标检测：代码实现与YOLOv7模型训练

目标检测：代码实现与YOLOv7模型训练

1. 目标检测代码实现

1.1 infer_object() 函数

infer_object() 函数用于推断 run_inference_for_single_image() 函数返回的 output_dict ，在图像中检测到的每个对象周围绘制边界框，并用类名和分数标记对象，最后将结果保存到输出目录。以下是该函数的代码：

def infer_object(model, image_path):
    # Read the image using openCV and create an image numpy
    # The final output image with boxes and labels on it.
    imagename = os.path.basename(image_path)

    image_np = cv2.imread(os.path.abspath(image_path))
    # Actual detection.
    output_dict = run_inference_for_single_image(model, image_np)

    # Visualization of the results of a detection.
    for i in range(output_dict['detection_classes'].size):

        box = output_dict['dete