A Python Flask API built in one week as a coding challenge. It analyzes uploaded images of meteorites to estimate their mineral composition, operational mining risk, and surface topology. Because I didn't have much time, some values in calculations were taken randomly and may not be so accurate.

How it works: Feature Extraction: The API uses OpenCV and Otsu's thresholding to separate the meteorite from the background. It calculates color averages (RGB), albedo, and surface depth (standard deviation).

Machine Learning Inference: The extracted stats are passed into a pre-trained XGBoost model to predict concentrations of Nickel, Iridium, and Gold. If the model file is missing or fails, the code uses a hardcoded fallback formula.

Operations Logic: The API takes UI inputs (orbit type and drill material) and applies multipliers to calculate the final value per ton, mission risk, and expected tool life.

Topology Mapping: It uses OpenCV's SimpleBlobDetector to find surface anomalies and creates a 128x128 pixel grid. This grid is smoothed with a Gaussian blur and returned as a list of lists to draw 3D terrain on a frontend.

Example of json: { "composition": { "nickel_pct": 10.5, "iridium_ppm": 3.2, "gold_ppm": 0.8 }, "economics": { "value_usd_per_ton": 2500.0 }, "engineering": { "hardness_score": 80.0, "mission_risk": 45.0, "tool_life_hrs": 500.0, "abrasive_index": 12.0 }, "visual_params": { "albedo": 0.25, "displacement_scale": 0.5, "topology": [[0, 0, ...], [...]] } }