Using camera recordings, the current work seeks to develop an automated method for generating virtual warfighter models to predict blast exposure in weapon training scenarios. The key question is whether we can accelerate the process of creating these virtual service member models for rapid exposure estimation. This work uses the latest machine learning-based tools for 3D human pose estimation from a single camera.
These tools allow us to extract the position and posture of each person in an image, streamlining the process of simulating a blast exposure. Using other sensors is difficult because service members in training don't have time to put on many different sensors. However, a camera can easily record a military training session, so our work leverages that modality to overcome the limitations of other types of sensors.
Our BOP Tool is the first computational tool to predict blast overpressure on service members using fast-running models optimized with field pressure and sensor data. It systematically replicate service members'posture and position during weapon firing, aiming to accurately estimate overpressure on different anatomical engines of the service member. We aim to enhance and transition the current blast overpressure tool into a real-time blast overpressure monitoring product for accumulator exposure.
We will also work towards correlating this dose to TBA response for injury risk assessment. Begin by clicking the BOP Tool scene module button to open the BOP Tool scene module from the BOP Tool interface. Click on scenario definition and then click on the scenario detail tab.
To begin reading and processing the image data, click the pose tool import button in the BOP Tool scene module, A popup window asking the user to select the relevant image or video will appear on the screen. Using mouse operations, navigate to the folder and select the relevant image or video. Once an image is selected, click open in the window popup, Use the text box under scenario name to choose a name for the scenario.
Under the name field, in the weapon definition, choose a custom name. Select the appropriate weapon class, such as heavy mortar, from the list of options in the dropdown menu. Next, select the weapon from the list of choices in the dropdown.
Using the dropdown options, select the ammunition shell full range training round for the chosen weapon system. Select no shooter under the anthropometry, posture, helmet, and protective armor fields under the shooter definition from the dropdown options. After choosing these options, the corresponding blast kernel will be automatically loaded into the GUI under the charges subtab.
Now, click on the service members tab to control the position and orientation of the service members imported from the image data. Use the X, Y, Z and rotation options to adjust the position and orientation. Use the name field to assign custom names to virtual service member models imported into the GUI.
The default names for the automatically imported models are S1, S2, S3, S4, and S5.Then, use the delete button to remove S2, S4, and S5.Navigate to the sensors tab. Click on add sensor to add a new virtual sensor, then choose a custom name under the name field. Select the type as virtual by clicking the dropdown under the sensor type field.
Edit the text boxes under the X, Y, and Z fields to choose the sensor's location. For the scenario discussed here, create four different sensors at four different locations. Choose V2, V3, and V4 as the names for the additional sensors.
Leave the rotation value as zero. To save the weapon training scene, click the save scenario button in the GUI at the top. Then, click the run scenario button at the bottom to run the blast overpressure simulation for the M120 weapon training scene.
The progress bar at the bottom shows the simulation's progress. Navigate to the model view tab to examine the blast overpressure exposure simulation. Click the current button to load the completed simulation into the visualization window.
Once the simulation is loaded, visualize it using the play button at the bottom of the screen. After reviewing the simulation, click on the blast load metrics tab, then, click the current button to load the overpressure plots at the virtual sensor locations. Finally, navigate to the dropdown series to display and select, checking the box for the corresponding sensor to plot the overpressure recorded at that virtual sensor.
The virtual training scene created from the image data accurately represented the image data. The BOP simulations provided a detailed visualization of the overpressure loads on the virtual service member in the scene at different instance over time.
