Fall of Shot Feedback

Updated 7/05/15 - New videos cleared for release.

Deep Vision's technology provides instantaneous fall-of-shot statistics for small arms training.

In a project supported by the UK Ministry of Defence (MoD) through the Centre for Defence Enterprise (CDE) Enduring Challenge programme, Deep Vision has exploited its state-of-the-art machine perception technology to develop a system that detects the point of impact of shots during live fire training exercises. Using passive (COTS) camera technology, Deep Vision can track the training target, detect the impact of the shots on the water surface, and measure the accuracy of the shot relative to the intended target.

Fall-of-shot feedback and statics are relayed instantaneously to the gunner for shot correction. Deep Vision's technology promotes the optimisation of live fire training exercises through dynamic learning, improved shooter accuracy, and economic efficiency.

Technology Highlights
  • Runtime training target designation
  • Automatic training target tracking
  • Passive fall-of-shot detection
  • Quantification of miss-distance
  • Real-time shooter performance feedback/statistics
  • Quantification of shooter bias
  • Robust to environmental conditions
Technology Details
  • Maintenance of a 30+ fps detection/tracking rate
  • Implemented in C/C++
  • Developed in GNU/Linux, but operating system agnostic
  • Small memory footprint
  • Customisable to meet current and future product/programme requirements
  • Support for mobile devices, such as the Android operating system
  • Support for embedded platforms such as: Gumstix, SheevaPlug, Raspberry Pi, etc.

Videos

Designation and tracking of a training target and detection of fall-of-shot in real-time. Basic performance statistics are enumerated, including the number of missed shots and the shooter bias. Designation and tracking of a training target, detection of fall-of-shot, and presentation of operator statistics. The relative miss-distance between the shot and the target is displayed. Real-time performance statistics are shown.
  • Target (80x80 px) in motion
  • Sea State disturbances
  • Prominent wake
  • Camera panning
  • Isolated, infrequent shots
  • Runtime shooter performance: missed shots, shooter bias
Designation and tracking of a training target and detection of fall-of-shot in real-time. Basic performance statistics are enumerated, including the number of missed shots and the shooter bias. Designation and tracking of a training target, detection of fall-of-shot, and presentation of operator statistics. The relative miss-distance between the shot and the target is displayed. Real-time performance statistics are shown.
  • Target (35x35 px) in motion
  • Sea State disturbances
  • Prominent wake
  • Camera panning
  • High contrast imagery
  • High shot density
  • Runtime shooter performance: missed shots, shooter bias
Designation and tracking of a training target and detection of fall-of-shot in real-time. Basic performance statistics are enumerated, including the number of missed shots and the shooter bias. Designation and tracking of a training target, detection of fall-of-shot, and presentation of operator statistics. The relative miss-distance between the shot and the target is displayed. Real-time performance statistics are shown.
  • Target (10x10 px) static in scene
  • Calm sea state
  • Prominent wake
  • No camera motion
  • Frequent target occlusion
  • High shot density
  • Runtime shooter performance: missed shots, shooter bias
Designation and tracking of a training target and detection of fall-of-shot in real-time. Basic performance statistics are enumerated, including the number of missed shots and the shooter bias. Designation and tracking of a training target, detection of fall-of-shot, and presentation of operator statistics. The relative miss-distance between the shot and the target is displayed. Real-time performance statistics are shown, as well as a cumulative view of missed shot frequency for the exercise.
  • Target (12px to 200px) in motion
  • Calm sea state
  • Prominent wake
  • Rapid Camera motion
  • Rapid Target motion
  • High shot density
  • Runtime shooter performance: missed shots, shooter bias