Subversion Repositories seema-scanner

In order to make modifications and test (e.g. change some parameters in the reconstruction process), the SVN repository should be checked out/compiled in a seperate user account. The software is linked against Qt, OpenCV and PCL in the current Ubuntu LTS release. Additional PPAs may be required, see the README file in the software repository. This ensures easy compilation and install. 

In order to make modifications and test (e.g. change some parameters in the reconstruction process), the SVN repository should be checked out/compiled in a seperate user account. The software is linked against the default versions of Qt, OpenCV and PCL in the current Ubuntu LTS release. This ensures easy compilation and install. 

In contrast to the SeeMa-Scanner located in the Image Lab, the Traveling SeeMa-Scanner first has to be assembled from the single parts shown in Figure \dolmes{XXX (Figure of box)}. Figure \dolmes{XXX} illustrates the final setup of the Traveling SeeMa-Scanner. Take the following points into consideration:

In contrast to the SeeMa-Scanner located in the Image Lab, the Traveling SeeMa-Scanner first has to be assembled from the parts stored in a large box. Figures \ref{fig:setup0} and \ref{fig:setup1} illustrate the final setup of the Traveling SeeMa-Scanner. Take the following points into consideration:

	\item Choose a black, non-shiny background (e.g. provided black fabric). Cover (or paint black) any shiny objects in the scan area such as screws, stands, etc. This will ensure that only the object of interest is scanned.

	\item Choose a black, non-shiny background (e.g. provided black fabric). Cover (or even paint black) any shiny objects in the scan area such as screws, stands, etc. This will ensure that only the object of interest is scanned.

	\item How to choose the distance between the projector and the circular rotation plate? This distance is dependent on the object size. Make sure the object is inside the field of view of cameras and projector for any rotation angle. \dolmes{Would you rather want a small distance?}

	\item How to choose the distance between the projector and the circular rotation plate? This distance is dependent on the object size. Make sure the object is inside the field of view of both cameras and the projector for any rotation angle.

\section{Calibration}

\section{Calibration}\label{sec:calibration}

	\item The GUI application is started on the scanner computer \dolmes{name/icon?}. Some software settings can be altered through the ''File $\rightarrow$ Preference'' menu, if necessary (the GUI needs to be restarted after altering these settings).

	\item The GUI application "SeeMaLab 3D Scanner" is started on the scanner computer by clicking on the icon \raisebox{-\mydepth}{\fbox{\includegraphics[height=\myheight]{images/icon1.png}}}. Some software settings can be altered through the ''SMScanner $\rightarrow$ Preferences'' menu, if necessary (see Figure \ref{fig:preferences_menu}). For the calibration part, choose the "Calibration" tab.

	\item Ensure that both the cameras and the projector are in focus at the middle of the rotation stage. \dolmes{You can check this by looking on the camera images in the GUI, and by holding your hand over the rotation plate for the projector.}

	\item Additionally, ensure that both cameras are in focus at the middle of the rotation stage. To do so, look at the camera images in the GUI, where you can zoom in by turning the mouse wheel, and check whether the projected pattern is sharp on the object (see Figure \ref{fig:projected_pattern}). If needed, the cameras can be focused by turning the corresponding focus ring, which is located in front of the camera's lens.

	\item Make sure the size of the calibration plate fits the object. \dolmes{where/how to adjust the parameters of the calibration plate?}

	\item Make sure the size of the calibration target (checkerboard plate) fits the object. Then, press ''SMScanner $\rightarrow$ Preferences'' and adjust the parameters of the calibration target (Calibration pattern; and Size). In order to determine the calibration pattern of the checkerboard plate, do NOT count the number of squares in horizontal and vertical direction, but the number of inner edges (saddle points). The parameter Size is given by the side length of a square. In Figure \ref{fig:calibration_target}, the calibration pattern is given by $13\times 22$ and the side length by $15$ mm.

	\item Position the calibration target on the circular rotation plate parallel to the projector -- make sure the rotation axis approximately intersects the center of the calibration target --, and inside the field of view of cameras and projector. White light will be provided from the projector for guidance. The GUI will show as shown on figure \ref{fig:calibration0}.

	\item Position the calibration target on the circular rotation plate parallel to the projector -- make sure the rotation axis approximately intersects the center of the calibration target --, and inside the field of view of both cameras and the projector. You can for example use four screws to mount the target on the rotation plate. White light will be provided from the projector for guidance. The GUI should look similar to the one in Figure \ref{fig:calibration0}.

	\item Optimally, the camera images in the GUI show a rather grayish calibration plate, and the background is totally black. If any pixels of the calibration plate are completely white, there is too much light. There are two options for adjusting the light:

	\item\label{item:light} Optimally, the camera images in the GUI show a rather grayish calibration plate, and the background is totally black. If any pixels of the calibration plate are completely white, there is too much light. There are two options for adjusting the light:

		\item Adjust the \dolmes{camera/projector?} shutter time by \dolmes{how?}. The shutter time has to be a multiple of 16.666 milliseconds per image: 16.666, 33.333, 50.000, 66.666, 83.333, 100.000, 116.666, 133.333, etc (type \dolmes{how many?} digits after the decimal point!). This is due to the fact that the projector has a \dolmes{frame rate?} of 60 images per second, which corresponds to a \dolmes{projection time?} of 1/60 seconds per image, i.e. 16.666 milliseconds per image. The larger the shutter time, the more light reaches the image plane, and vice versa. Note that changing the shutter time does not affect calibration!

		\item Adjust the shutter time in the ''SMScanner $\rightarrow$ Preferences'' menu. The shutter time has to be a multiple of 16.666 milliseconds per image: 16.666, 33.333, 50.000, 66.666, 83.333, 99.996, 116.666, 133.333, etc (important: type 3 digits after the decimal point!). This is due to the fact that the projector has a frame rate of 60 images per second, which corresponds to a projection time of 1/60 seconds per image, i.e. 16.666 milliseconds per image. The larger the shutter time, the more light reaches the image plane, and vice versa. Note that changing the shutter time does not affect calibration, i.e. the shutter time can also be changed after having done a calibration, without having to do a recalibration!

	\item A number of calibration sets need to be acquired. The minimum is 3 sets, and more is beneficial. The calibration pattern needs to be fully visible and equally bright in both cameras. The viewing angle must not be too shallow. Press ''batch acquisition'' in order to acquire a reasonable number of calibration sets using default parameters. \dolmes{Figure \ref{fig:calibration0} reveals that default acquisition is at angles $330^\circ, 335^\circ, \dots, 30^\circ$}. %The present ''batch acquisition'' gives a reasonable number of calibration sets.

	\item\label{item:batch1} A number of calibration sets need to be acquired. The minimum is 3 sets, and more is beneficial. The calibration pattern needs to be fully visible and equally bright in both cameras. The viewing angle must not be too shallow. Press ''Batch acquisition'' in order to acquire a reasonable number of calibration sets using default parameters. Figure \ref{fig:calibration0} reveals that default acquisition is at angles $340^\circ, 338^\circ, \dots, 20^\circ$. %The present ''batch acquisition'' gives a reasonable number of calibration sets.

	\item After acquisition, individual calibration sets can be re-examined. Calibration parameters are automatically determined by clicking the ''Calibrate'' button. This procedure can take up to a few minutes. The terminal output will show recalibration errors, which measure the quality of calibration. 

%untick all the images acquired in calibration steps \ref{item:batch2} and \ref{item:batch3}. By doing so, you make sure that only the images acquired in step \ref{item:batch1} are used for determination of the rotation axis, whereas all images are used for camera calibration. By clicking the ''Calibrate'' button, calibration parameters are then automatically determined. This procedure can take up to a few minutes. The terminal output will show recalibration errors, which measure the quality of calibration.

	\item A successful calibration goes along with a colorful pattern on the calibration target as shown in Figure \dolmes{XXX}. \dolmes{If the calibration fails for a calibration set, it is automatically ignored. Thus, it does not matter if the calibration is not successful for a few sets.} In addition,the calibration result can be examined by changing to the ''Point Clouds'' tab in the GUI (see Figure \ref{fig:pointclouds0}). Left and right cameras are represented by coloured coordinate systems (the viewing direction is the positive z-axis, y points down, x to the right). The rotation axis, as determined by the calibration procedure is shown as a white line section. 

	\item A successful calibration goes along with a colorful pattern on the calibration target as shown in Figure \ref{fig:successful_calibration}. If the calibration fails for a calibration set, it is automatically ignored. Thus, it does not matter if the calibration is not successful for a few sets. In addition, the calibration result can be examined by changing to the ''Point Clouds'' tab in the GUI (see Figure \ref{fig:pointclouds0}). Left and right cameras are represented by coloured coordinate systems (the viewing direction is the positive z-axis, y points down, x to the right). The rotation axis, as determined by the calibration procedure is shown as a white line.

\section{Making a 360 degree scan}

\section{Making a 360 degree scan}\label{sec:scan}

Image acquisition consists of projecting a sequence of patterns onto the object, which are then converted to depth values by means of the specific algorithm.\\

Image acquisition consists of projecting a sequence of patterns onto the object, which are then converted to depth values by means of the specific algorithm. You can choose among different pattern modes in the ''SMScanner $\rightarrow$ Preferences'' menu:

Depending on the surface complexity (blind spots, holes, details, etc.), multiple $360^\circ$ scans may be necessary. In that case, the following procedure is done multiple times with the object in different orientations (poses). Also consider to change the rotation angle in order to obtain a better result.

Depending on the surface complexity of the scan object (blind spots, holes, details, etc.), multiple $360^\circ$ scans may be necessary. In that case, the following procedure is done multiple times with the object in different orientations (poses) in order to cover the whole surface, and capture all details. Consider to change the rotation angle for a better result. In order to obtain a good quality scan, the number of poses, as well as the poses and rotation angles used for the scanning have to be carefully investigated.

	\item Choose the ''Capture'' tab in the GUI -- see figure \ref{fig:capture0} for an illustration.

	\item Choose the ''Capture'' tab in the GUI.

	\item Check the \dolmes{light conditions}: Again, the object should appear grayish with a completely black background. \dolmes{If necessary, adjust the light conditions}, preferably by changing the shutter time as described in the calibration part, since re-calibration is not needed.

	\item Check the light conditions: Again, the object should appear grayish with a completely black background (see Figure \ref{fig:light}). If necessary, adjust the light conditions, preferably by changing the shutter time as described in the calibration part (see Section \ref{sec:calibration}, step \ref{item:light}), since recalibration is not needed.

	\item Press ''Single Capture'' or ''Batch Capture'' in the GUI.

	\item Press ''Single Capture'' or ''Batch Capture'' in the GUI in order to scan the object.

	\item Sequences of patterns are projected onto the object \dolmes{different methods? where to choose?}. The captured images can be reviewed, and one or multiple captured sequences reconstructed using the ''Reconstruct'' button. 

	\item Sequences of patterns are projected onto the object, and images are acquired. The captured images can be reviewed by clicking on the frames (see Figure \ref{fig:capture1}). Captured sequences are automatically reconstructed, where the name of a reconstructed sequence appears black, otherwise it is grey (see Figure \ref{fig:capture0}).

	\item The results will show up in the ''Points Clouds'' tab \dolmes{how to zoom in - a button has to be pressed?}. Single point clouds can be shown or hidden, see figure \ref{fig:pointclouds1}.

	\item The results can be investigated by choosing the ''Points Clouds'' tab in the GUI (see Figure \ref{fig:pointclouds1}). In order to zoom in at a specific point, hover the mouse over this point and press F. Single point clouds can be shown (ticked), or hidden (unticked). % click on a point + press F

	\item All data can be exported from the GUI program by means of the top bar menus. By exporting the point clouds into a folder, a \texttt{*.aln} is stored alongside these, which contains pose information in global coordinate space, which aligns the points clouds correctly and relative to each other.

	\item All data can be exported from the GUI program by means of the top bar menus. By exporting the point clouds into a folder (''Point Clouds $\rightarrow$ Export Point Clouds''), a \texttt{*.aln} is stored alongside these, which contains pose information in global coordinate space, which aligns the points clouds correctly and relative to each other. The captured images can be exported by either pressing ''Capture $\rightarrow$ Export Sequences'' (whole sequence) or ''Capture $\rightarrow$ Export White Frames'' (no images showing projected patterns on the object). 

		\includegraphics[width=.7\textwidth]{calibration0.png}

		\includegraphics[width=.7\textwidth]{images/calibration_tab.png}

	\caption{The GUI showing the ''Calibration'' tab.}

	\caption{GUI showing the ''Calibration'' tab}

		\includegraphics[width=.7\textwidth]{pointclouds0.png}

		\includegraphics[width=.7\textwidth]{images/pointclouds0.png}

	\caption{GUI showing the result of calibration in the ''Point Clouds'' tab.}

	\caption{GUI showing the calibration result in the ''Point Clouds'' tab}

		\includegraphics[width=.7\textwidth]{capture0.png}

		\includegraphics[width=.7\textwidth]{images/capture0.png}

	\caption{The ''Capture'' tab in the GUI.}

	\caption{GUI showing the ''Capture'' tab}

		\includegraphics[width=.7\textwidth]{pointclouds1.png}

		\includegraphics[width=.7\textwidth]{images/pointclouds1.png}

	\caption{''Point Clouds'' tab with reconstructed point clouds.}

	\caption{''Point Clouds'' tab with reconstructed point clouds}