Radiation Heatmap & Image Processing Engine
Custom convolution kernels handle filtering and denoise, then intensity maps into HSV so magnitude reads as colour. Built from matrix operations rather than library calls — the filters are written out, not imported.
- Stack
- MATLABImage ProcessingLinear AlgebraSignal Processing
- Highlights
- Custom convolution kernels for filter and denoise passes
- Intensity mapped to HSV heatmaps
- Matrix-level implementation without black-box library filters
The problem
Raw sensor frames are noisy and hard to read as numbers in a table. The goal was a pipeline that cleans the signal and makes magnitude obvious visually.
Approach
Write the filters. Custom convolution kernels for filtering and denoise, implemented as matrix operations so the math is visible and inspectable.
Colour as magnitude. After cleanup, intensity maps into HSV space so the heatmap carries the story instead of a spreadsheet of values.
Outcome
A small, honest image-processing exercise: understanding the kernels mattered more than shipping a polished UI.