Now if you install a few Raspberry Pi's with a cheap GPS and a cheap microphone you could likely use my sound localization project and perform very accurate loc

Now if you install a few Raspberry Pi's with a cheap GPS and a cheap microphone you could likely use my sound localization project and perform very accurate localization and know for sure in future incidents: https://github.com/hcfman/sbts-aru https://github.com/hcfman/sbts-aru Nice to see you using sound calculations if your analysis. I don't know how hot it was at the time but a simple heuristic to adjust the speed of sound might make a (very slight) difference. Speed approx = 331.3 + 0.6 * temp_in_celsius But if you can place at least 3 (Technically 4x, but the best of two solutions will be chosen if you only have 3x) mics around the place, I'm pretty sure that the sound of the explosions could be heard for miles and your should have pretty darn good localization accuracy and be independant from webcam from solely relying on webcam images.