*3.2. Data Collection Procedure*

The data collection procedure was performed similar to the works in the literature [1–3]. For both concentrated and distributed loading scenarios, first, loads were applied at each position as depicted in Figure 4 separately, and then depth images of the wing were acquired. For example, a load of 2.45 N was applied on position 1, while no other load was applied at any other location, and then the data was collected. Afterward, the next load of 4.9 N was applied at the same position, and the data were recorded. This procedure was repeated for all other locations in the same manner until data from all of the positions with all of the different load magnitudes were recorded. The acquired depth images by Kinect V1 are shown in Figure 7, in which the pixel values correspond to the actual measured distance in mm. Because the images were captured at one meter distance with a resolution of 640 × 480 pixels, the captured scene encompassed much more information than required; therefore, the images were cropped to include only the UAV wing, and the size of the acquired image was reduced to 247 × 166 pixels. It should be noted that this was only done in this case and, if the whole wing encompasses the image, then there is no need for cropping the image. Moreover, any depth values above 2000 mm and below 800 mm were changed to zero in order to get rid of redundancies, as shown in the right image

of Figure 7. Moreover, since the gap between the depth values of the wing and its surrounding were very large, the color distribution resulted in a binarized representation.

**Figure 7.** (**Left**) Acquired and (**Right**) Extracted depth image of the composite UAV wing.
