{
  "fetched_at": "2026-10-19T09:37:08Z",
  "source_url": "https://www.kaggle.com/competitions/ayb-ceng463-machine-learning-project-fall-2022/data",
  "pages": [
    {
      "name": "data-description",
      "content": "Gibraltar is selected as a study area as there is wiwide variety of land cover types. ",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "# CENG 463 Project Fall 2022\nClimate change has become a serious threat to humanity. Global land cover maps serve as an important means to tackle this problem. ESA has produced Global Land Cover Map from Sentinel-2 data. \nWe will generate land cover maps for unseen data by using a limited number of training samples. Gibraltar is selected as a study area as there is a wide variety of land cover types. \n## Data:\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2285546%2F61185976705294ac2f34dd74ddde425d%2FTrainingData.jpg?generation=1670604296605320&alt=media)\n\nData Link: https://drive.google.com/drive/folders/1diOV3oaKTsXul3SB8YSWpEQ8AvtZO-Uk?usp=sharing \nTest Data will be provided later. \nData is UInt16, Class labels are Uint8. \nTraining data: S2A_MSIL1C_20220516_TrainingData.tif \nTraining labels: S2A_MSIL1C_20220516_Train_GT.tif \n\n## Class List: \n ![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2285546%2F01ce381c9ec97efd5bf5e6e2a2a4671e%2FCodes.png?generation=1670604363281149&alt=media)\n\nReference: https://esa-worldcover.s3.eu-central-1.amazonaws.com/v200/2021/docs/WorldCover_PUM_V2.0.pdf \nNote that some of the classes may not be present in the data set. You may ignore classes that are not present in both datasets. \n## Methods:\nYou are limited to methods that are covered in the course however you may perform feature engineering such as selecting bands or taking ratios of bands. Methods such as neural networks or deep learning should not be used. \n## Software: \nYou may only use SKLearn, OpenCV, GDAL (Data Read), Numpy, and Matplotlib (Those used in your lab sessions.)\n## Deadline: \nThe week after the end of the final exams. \n## Report: \nYou should prepare a report which includes your implementation details, and a link to your GitHub repo which includes codes. In addition, you may prepare a report in IEEE Paper format such as RAST2023 conference. \nYour report should include details about your project such as the number of training samples, how you selected a method, and how it performs with the provided data. You should search literature briefly from Google Scholar (spend the 1st week’s effort).\nPresentation & Competition: \nThere will be project presentations by project teams. There will be a competition based on your classification score, report, and presentation. \n## Teams: \nYou can team up to 2 students. \n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "F-Score will be used as the metric.\n50% of the test data will be used in public, and 50% will be used to evaluate in the private leaderboard.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "You may only use SKLearn, OpenCV, GDAL (Data Read), Numpy, Matplotlib (Those used in your lab sessions.)",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}