# Challenge 1 — Data

来源：https://www.kaggle.com/competitions/challenge-1/data

采集时间：2026-10-20T01:30:09Z

# Training
Training data for your models : the Belgian TS dataset, available at https://btsd.ethz.ch/shareddata/
You can start with the dataset « BelgiumTS for Classification (cropped images) »
Please do not train on data from other european countries.


# Evaluation
Predict the class (aka category) of the pictures collected in directory eval_kaggle1.
Write your predictions in a .csv file with the following format:

```
Id,Category
1,13
2,1
3,42
...
202,55
```

where the first number 'Id' is the name of the picture and the second number is the class (aka category) computed by your algorithm, i.e. a number between 0 and 61.
An example of a submission file with correct format is given: see submission_file.csv.
Note that the categories of the file 'submission_file.csv' have been randomly generated.

# Metric

This competition is evaluated on the categorization accuracy of your predictions (the percentage of images you get correct).