pytorch opencv pil图像预处理比较

opencv的：

  

   

    

     

    
    

     

                      img = cv2.imdecode(np.fromfile(label_file.replace("label.csv","")+img_path, dtype=np.uint8), 1)
     
    
   

    

     

    
    

     
                print(label_file.replace("label.csv","")+img_path)
     
    
   

    

     

    
    

     
                if img is None:
     
    
   

    

     

    
    

     
                    print('file is none',label_file.replace("label.csv","")+img_path)
     
    
   

    

     

    
    

     

                          df_train.drop(df_train[df_train.img_path == img_path].index, inplace=True)
     
    
   

    

     

    
    

     

                          df_result.drop(df_train[df_train.img_path == img_path].index, inplace=True)
     
    
   

    

     

    
    

     

                          df_train.to_csv(label_file, index=False)
     
    
   

    

     

    
    

     
                    continue
     
    
   

    

     

    
    

     

                      img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
     
    
   

    

     

    
    

     

                      t_h, t_w = img.shape[:2]
     
    
   

    

     

    
    

     

                      a_w = max(t_w, t_h)
     
    
   

    

     

    
    

     

                      img_b = np.zeros((a_w, a_w, 3), dtype=np.uint8)
     
    
   

    

     

    
    

     
                if t_h > t_
     
    
  
 

文章来源: blog.csdn.net，作者：AI视觉网奇，版权归原作者所有，如需转载，请联系作者。

原文链接：blog.csdn.net/jacke121/article/details/125942704