In photogrammetry, a set of high-quality images of the test object is crucial for reconstructing the model. Image capture is mostly done manually, which is time consuming and difficult to achieve good repeatability. This master's thesis discusses the development of a mechatronic system for automatic image capturing of small objects for photogrammetry purposes. This approach increases the speed of image capture and ensures high precision and good repeatability. The system is composed of parts mostly made by 3D printing, with kinematics achieved using stepper motors. An industrial camera is used for image capture, and lighting is also controlled. An Arduino microcontroller is used, which communicates with a Python script on a computer.
We found that the system is reliable and captures images much faster than manual capture. It was observed that a larger number of images positively impacts the quality of the reconstructed models, with the texture of the objects being a very important factor for quality. The reconstruction of small objects is less accurate due to the low resolution of the camera.
|
