将PPM格式图片转化为JPG格式

做图像识别的时候数据集常常是ppm格式的，虽然不影响建模训练，但是我们电脑往往不支持ppm格式的图像展示。

比如到做交通标志识别的时候用到的BelgiumTS交通数据集或者德国GTSRB数据集

下载后得到都是ppm格式的图像。

格式转化代码

from PIL import Imageimport osinput_train_path = r"E:\test\BelgiumTSC_Training\Training"output_train_path = r"E:\BelgiumTSC_JPG\Train_data"input_test_path = r"E:\test\BelgiumTSC_Testing\Testing"output_test_path = r"E:\BelgiumTSC_JPG\Test_data"def batch_image(in_dir, out_dir):if not os.path.exists(out_dir):print(out_dir, 'is not existed.')os.mkdir(out_dir)if not os.path.exists(in_dir):print(in_dir, 'is not existed.')return -1directories = [d for d in os.listdir(in_dir) if os.path.isdir(os.path.join(in_dir, d))]for d in directories:# 每一类的路径label_directory = os.path.join(in_dir, d)new_directory = os.path.join(out_dir, d)out_folder = os.path.exists(out_dir + d)if not out_folder:os.mkdir(new_directory)file_names = [os.path.join(label_directory, f) for f in os.listdir(label_directory) if f.endswith(".ppm")]# file_names is every photo which is end with ".ppm"count = 0for files in file_names:file_path, extfilename = os.path.split(files)filename, extname = os.path.splitext(extfilename)out_file = filename + '.jpg'# print(filepath,',',filename, ',', out_file)im = Image.open(files)new_path = os.path.join(new_directory, out_file)print(count, ',', new_path)count = count + 1im.save(new_path)if __name__ == '__main__':batch_image(input_test_path, output_test_path)batch_image(input_train_path, output_train_path)

转化后会保留原有的两级目录，方便建模训练。