The aim of the project is to develop a sterilizable endoscopic camera calibration device that cam be used in operating theaters by clinicians. Medical imaging methods such as real time tracking of the human cardiovascular chart, currently suffer from optical distortion effects that could be mitigated by a standardized calibration process. This device can serve as a quick and practical tool for the peer operative calibration of endoscopes used in fetal surgery procedures, such as twin to twin transfusion syndrome.
Before sandblasting use a ruler to draw a 40 by 40 millimeter square onto a 1.2 millimeter thick, 316 stainless steel sheet. Then use a manual metal cutter to cut out the square and use a file to round the corners and sides of the sample. Next place the steel onto a one to two centimeter thick block that is slightly larger than the square sample.
Use one vice to secure the sample onto the block and another to hold the sample tightly by the edge of the block during the sanding. Then move the entire assembly into an internal blast chamber turn on the compressed airflow, and tightly seal the chamber. Then don a pair of safety goggles and position a blast gun perpendicular to and at least four to five centimeters away from the metal surface.
Next apply the foot control to sandblast the sample. Then sandblast the other side as just demonstrated as appropriate. For laser patterning of the sample first design a pattern of asymmetric circles and use the appropriate design software to prepare a drawing exchange format file of the design.
Import the file into the laser cutting software and set the background etching and the etching pattern perimeters. Go to the laser tab of the smartest software. When you choose stainless steel as the material predefined perimeters will appear.
Adjust the power, shot frequency, scan speed, and number of passes perimeters accordingly. Adjust the X and Y position of the pattern to be etched if required. To recognize the pattern using end accounts of squares the color of the block and background should be optimized.
The correct performance and the perimeter setting of the laser should also be checked before beginning the patterning. Next, put the sample on the working platform, focus the surface of the sample using the Z axis command window to change the height of the laser. Check that the start test option is activated on the software, and switch on the laser diode on the laser panel, and then adjust the height of the laser.
When the test spot coincides with the diode laser spot the focus of the laser is exactly on the surface. Use the software to align the cutting pattern. Dip the pattern sample into alcohol to clean the surface.
Then wrap the sample in a sterilization package for autoclaving. To calibrate the endoscope first install the Endocal endoscope calibration software package provided on GitHub. Then place the calibration target in a steril fluid container, such as a gallipot.
Fill the container with the target fluid and adjust the zoom and sharpness of the endoscope as desired. Next immerse the endoscope in the fluid, holding the endoscope at a distance from the calibration from the calibration target, similar to the distance at which the endoscope will later be used. Now launch the calibration application and begin the camera acquisition, moving the tip of the endoscope slightly for different views while keeping the whole calibration pattern in the view of the camera.
Acquire the minimum number of endoscopic camera views required for the calibration, as indicated in the Endocal window. This sterilizable calibration target was created by using an asymmetrical circle pattern onto a sandblasted stainless steel sheet as illustrated in these images. For the pattern to have a consistent color in the end the distance between the lines should be less than the width of the laser beam.
Here an exemplary set up of this calibration target in action with an endoscope is shown. The fabricated calibration target allows detection of the circular blobs in the endoscopic video stream with open CV.The locations of which are then sorted into predefined asymmetric circular grid geometry and subsequently used for estimating the optical distortion perimeters. Optical distortions can be corrected using the estimated perimeters.
For example, in this representative endoscopic view of a rectangular chess board pattern, the optical distortions make the lines appear as curves. After distortion correction however, the lines appear normal. This technique allows the affordable fabrication of a sterilizable device for endoscopic camera calibration from readily available medical grade material.
Building on the positive results from this work we are moving toward the development of a fully fledged ergonomic calibration device that will require only minimal user interaction. This production method can help answer key questions in the interpretive medical imaging field such as, how can the sterilability of the endoscopic camera be preserved during the calibration process? Our calibration device works with state of the art camera calibration software like Open CV, further reducing the complexity of integrating the device with many potential clinical software applications.
