---
license: cc-by-nc-3.0
tags:
- audio
- audio-classification
- military-audio
- drone-detection
- environmental-sound
pretty_name: 'Fusion MAD + ESC-50 + DroneAudioDataset '
size_categories:
- 1B<n<10B
---

# mad-esc50-drone-audio

Dataset audio fusionné à partir de 3 sources, normalisé pour la classification audio (colonnes : `audio`, `label`, `source_dataset`, `split`).

## Sources et attributions

- **MAD (Military Audio Dataset)** — Kim, Yoon & Jung, 2024, *Scientific Data*. CC BY 4.0.
  https://www.kaggle.com/datasets/junewookim/mad-dataset-military-audio-dataset
- **ESC-50** — Karol J. Piczak. CC BY-NC 3.0 (usage non commercial).
  https://github.com/karolpiczak/ESC-50
- **DroneAudioDataset** — Al-Emadi et al., IWCMC 2019. Citer le papier en cas d'usage recherche.
  https://github.com/saraalemadi/DroneAudioDataset

⚠️ Le sous-ensemble "unknown" de DroneAudioDataset réutilise des fichiers ESC-50 ; il a été exclu de la fusion pour éviter les doublons (voir notebook de construction).

59 classes :
   0. 0
   1. 1
   2. 2
   3. 3
   4. 4
   5. 5
   6. 6
   7. airplane
   8. bebop_1
   9. breathing
  10. brushing_teeth
  11. can_opening
  12. car_horn
  13. cat
  14. chainsaw
  15. chirping_birds
  16. church_bells
  17. clapping
  18. clock_alarm
  19. clock_tick
  20. coughing
  21. cow
  22. crackling_fire
  23. crickets
  24. crow
  25. crying_baby
  26. dog
  27. door_wood_creaks
  28. door_wood_knock
  29. drinking_sipping
  30. engine
  31. fireworks
  32. footsteps
  33. frog
  34. glass_breaking
  35. hand_saw
  36. helicopter
  37. hen
  38. insects
  39. keyboard_typing
  40. laughing
  41. membo_1
  42. mouse_click
  43. pig
  44. pouring_water
  45. rain
  46. rooster
  47. sea_waves
  48. sheep
  49. siren
  50. sneezing
  51. snoring
  52. thunderstorm
  53. toilet_flush
  54. train
  55. vacuum_cleaner
  56. washing_machine
  57. water_drops
  58. wind