为什么在尝试预处理图像时会出现重复的图像？

问：

我的 python 函数的结果不是我想要的。你能帮我解决这个问题吗？

我想将 pdf 转换为黑白图像，同时在转换过程中对其进行预处理。但是我遇到的问题是我得到了两个结果：一个是将 pdf 转换为没有预处理的图像，第二个（我想要的）是 image_processor 函数应用的图像预处理。我只想获得第二个结果，但以 pdf_to_images 函数中编写的形式。以下是有问题的两个函数。你有什么建议来解决这个问题？

MAX_SIZE_MB = 5

def check_image_size(image_path):
    file_size_MB = os.path.getsize(image_path) / (1024 * 1024)
    if file_size_MB > MAX_SIZE_MB:
        raise ValueError(f"The image size ({file_size_MB:.2f} MB) exceeds the allowed limit of {MAX_SIZE_MB} MB.")
    print(f"Image size: {file_size_MB:.2f} MB")



def processor(input_image_path, contrast_factor=1.5, brightness_factor=1.5, scale_factor=2):
   img_pil = Image.open(input_image_path)
   img_np = np.array(img_pil)
   if len(img_np.shape) == 3:
      img_cv = cv2.cvtColor(img_np, cv2.COLOR_RGB2GRAY)
   else:
      img_cv = img_np
   clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8))
   img_clahe = clahe.apply(img_cv)
   img_clahe_pil = Image.fromarray(img_clahe)
   enhancer_contrast = ImageEnhance.Contrast(img_clahe_pil)
   img_enhanced = enhancer_contrast.enhance(contrast_factor)
   enhancer_brightness = ImageEnhance.Brightness(img_enhanced)
   img_enhanced_bright = enhancer_brightness.enhance(brightness_factor)
   img_enhanced_bright_convert = img_enhanced_bright.convert('L')
   _, img_binary = cv2.threshold(np.array(img_enhanced_bright_convert), 0, 255, cv2.THRESH_BINARY + cv2.THRESH_OTSU)
   img_binary_pil = Image.fromarray(img_binary)
   new_width = int(img_binary_pil.width * scale_factor)
   new_height = int(img_binary_pil.height * scale_factor)
   img_upscaled = img_binary_pil.resize((new_width, new_height), Image.BICUBIC)
   output_path = input_image_path.replace('.jpeg', '_processed.png')
   img_upscaled.save(output_path, "PNG", quality=100)
   check_image_size(output_path)
   return output_path



def pdf_to_images(pdf_file, output_dir, dpi=600):
    doc = fitz.open(pdf_file)
    scale_factor = dpi / 72 
    zoom_matrix = fitz.Matrix(scale_factor, scale_factor)  

if not os.path.exists(output_dir):
    os.makedirs(output_dir)

for page in doc:
    pix = page.get_pixmap(alpha=False, matrix=zoom_matrix) 
    img = Image.frombytes("RGB", (pix.width, pix.height), pix.samples)
    img_gray = img.convert("L")  
    output_file = f"{output_dir}/page{page.number}.jpeg"
    # img_gray.filename = output_file  
    # img_gray.save(output_file, format='JPEG', quality=100, optimize=True)
    
    
    
    quality = 100
    while True:
        with open(output_file, 'wb') as f:
            img_gray.save(f, format='JPEG', quality=quality, optimize=True)
            size = f.tell()
        if size <= 1.5 * 1024 * 1024:  
            break
        if quality <= 10:  
            break
        quality -= 5  
        

    processed_image_path = processor(output_file)
    print(f"Processed image saved at {processed_image_path}")

这是我得到的结果：

2023-10-03 08:01:45.310 python[4319:178909] +[CATransaction synchronize] called within transaction
Image size: 1.07 MB
Processed image saved at /Users/Desktop/page0_processed.png
Image size: 0.97 MB
Processed image saved at /Users/Desktop/page1_processed.png
Image size: 0.78 MB
Processed image saved at /Users/Desktop/page2_processed.png
Image size: 1.00 MB
Processed image saved at /Users/Desktop/page3_processed.png
Image size: 0.85 MB
Processed image saved at /Users/Desktop/page4_processed.png
Image size: 0.66 MB
Processed image saved at /Users/Desktop/page5_processed.png
Image size: 0.81 MB
Processed image saved at /Users/Desktop/page6_processed.png
Image size: 0.55 MB
Processed image saved at /Users/Desktop/page7_processed.png
Image size: 0.45 MB
Processed image saved at /Users/Desktop/page8_processed.png
Image size: 0.44 MB
Processed image saved at /Users/Desktop/page9_processed.png
Image size: 0.43 MB
Processed image saved at /Users/Desktop/page10_processed.png
Image size: 0.38 MB
Processed image saved at /Users/Desktop/page11_processed.png
Image size: 0.37 MB
Processed image saved at /Users/Desktop/page12_processed.png