Predictive Modeling Contest
Description
Importance of Precise Crop Mapping for Precision Agriculture
Precise crop mapping is crucial for precision agriculture, which aims to optimize field-level management in crop farming. Accurate crop mapping helps in:
- Improved Resource Allocation: Identifying different crop types accurately allows for better allocation of resources such as water, fertilizers, and pesticides.
- Enhanced Yield Prediction: Understanding the spatial distribution of crops helps in predicting yields more accurately, facilitating better planning and market strategies.
- Environmental Sustainability: Precise mapping helps in monitoring and managing the health of crops, reducing waste, and minimizing the environmental impact of farming practices.
- Targeted Interventions: It enables the implementation of targeted interventions to address specific issues like disease outbreaks, nutrient deficiencies, or pest infestations.
In the context of the Senegalese groundnut basin, precise crop mapping is particularly important due to the region’s economic reliance on groundnut farming. Accurate crop maps can help in ensuring better crop management, improving yield predictions, and supporting sustainable agricultural practices, ultimately contributing to the overall food security and economic stability of the region.
Groundnut Basin of Senegal
The groundnut basin of Senegal is a significant agricultural region known for its extensive cultivation of groundnuts (peanuts). This region is pivotal to Senegal’s economy, providing substantial income and employment opportunities for the local population. The basin’s agricultural activities are highly dependent on seasonal rainfall and climatic conditions, making it crucial to monitor and manage crop health and productivity effectively.
Challenge Description
This data challenge focuses on developing a predictive model for crop mapping using remote sensing and machine learning techniques. The goal is to accurately predict crop types based on the values of optical bands from Sentinel-2 and calculated band indices derived from those optical bands. The dataset includes some optical bands and calculated band indices information collected from the Senegalese groundnut basin over a period from July 1, 2023, to October 31, 2023, in 15-day intervals as described in Table 1.
| Band Index | Description |
| Blue (BLUE) | Optical blue band |
| Green (GREEN) | Optical green band |
| Red (RED) | Optical red band |
| Near infrared (NIR) | Optical near infrared band |
| Wide Dynamic Range Vegetation Index (WDRVI) | Give the same information as NDVI but more important for crops with dense canopies and mature crops |
| Normalized Difference Vegetation Index (NDVI) | Used for monitoring the health and productivity of vegetation |
| Synthesized NDVI (SNDVI) | Give same information as NDVI but more accurate for crop growth events, and less accurate for crop decline and the sudden changes |
| chlorophyll index (CI) | Applied to calculate the total amount of chlorophyll in plants |
| MERIS terrestrial chlorophyll index (MTCI) | Provides better results for planophile canopies. |
| Plant Senescence Reflectance Index (PSRI) | Used for vegetation health monitoring, plant physiological stress detection and crop production, and yield analysis |
| Enhanced vegetation index (EVI) | Like NDVI, but uses the reflection region of blue light |
| Soil-Adjusted Vegetation Index (SAVI) | Used to minimize soil brightness influences using a soil-brightness correction factor |
Evaluation
To evaluate the accuracy of your predictive model for crop mapping, your submission file should contain predicted groundnut crop data for each pixel identified in the data description (Dataset_Sen_GB_Challenge_to_predict.csv). We have isolated a part of the dataset specifically for evaluating the accuracy of your model’s outputs. We will use the following metrics for evaluation:
- Confusion Matrix: This shows true positives, false positives, true negatives, and false negatives, providing a clear picture of the model’s errors.
- Overall Accuracy: This is the proportion of correctly predicted crop types out of the total predictions. It gives a basic performance measure but can be misleading with imbalanced datasets.
- Precision: The ratio of correctly predicted positive observations to the total predicted positives. It measures the accuracy of positive predictions.
- Recall (Sensitivity): The ratio of correctly predicted positive observations to all actual positives. It indicates the model’s ability to identify all relevant instances.
- F1-Score: The harmonic means of precision and recall, offering a balanced single metric, especially useful for imbalanced datasets.
Your submission should look like this:
| Pixel Id | Crop type |
| 00001 | Groundnut |
| 00002 | Groundnut |
| …… | Not Groundnut |
Rules
The solution must use publicly available and open-source packages.
- Only submission files containing prediction values for all pixels will be considered.
- Individuals can form teams to participate in this contest. However, you cannot submit your work as an individual while you are a member of a participating team.
- Only one submission is allowed. However, you can request a second submission and the ReSAKSS Data Challenge team will review it.
- If you are the winner, the ReSAKSS Data Challenge team will request you to submit your model for copyright and predictive values verification.
- If two solutions have the same scoring, the date and time of submissions will be considered to assess the winner.
- You agree to legally assign ownership of all rights of copyright of the winning solution code to the ReSAKSS Data Challenge while reserving the right to the solution code for non-commercial purposes while crediting the ReSAKSS Data Challenge.
- The ReSAKSS Data Challenge can change the rules at any time.
- Each candidate can only participate in one of the two IT contests.
Timeline
The IT Predictive Modeling category will open on July 12, 2024 and will close on August 20, 2024. Applications will be made through this dedicated ReSAKSS Challenge website.
Prize
Only 1st prize will be attributed to this IT Predictive Modeling contest. The Winner will receive $1500 USD, and will be awarded during the 2024 ReSAKSS Conference.
