Radar-Based Milk Spoilage Detection Dataset

Raw mmWave radar ADC dataset for milk spoilage detection (UNSW Honours thesis project). Contents- 16 raw DCA1000 ADC capture files (`.bin`, ~64 MiB total)- Radar/capture configs (`configs/Milk.mmwave.json`, `configs/Milk.setup.json`)- Reference scripts (`code/`) for dataset building and training/evaluation ConditionsThree acquisition groups: (1) Metal reflector + Light milk, (2) No reflector + Light milk, (3) No reflector + Full-cream milk. Time/state labels: 0H, 21H, 48H, 70H + No_Milk (empty control). Container: transparent plastic. HardwareTI AWR1843 (77 GHz) + DCA1000; 256 ADC samples/chirp; 64 loops/frame; 8 frames/capture; 40 ms frame periodicity. Labels & pHPer-file labels and measured pH values are provided in `LABELS.md` and `metadata.csv` (included in the release package). Please cite this Zenodo record (DOI) if used.