RADIS system development
Several software sub-systems are integrated into RADIS to obtain the required functionality . The design architecture for each sub-system is explained as follows :
Sub-system 1 : Camera video imaging and GPS coordinate capture , display and record A number of threads , namely the capture thread , graphic canvas thread ( per camera ), display thread , and recording thread , are designed to fulfil the required video image capture , display and recording functions of the sub-system . The sequential diagram representing the operation of each thread is shown in Figure 1 [ 3 , 4 ].
Plate 1 : Video image recording system
Figure 1 : Sequential diagram ( video image capturing , displaying and recording )
Using a similar concept , a number of threads were developed for the GPS capture ( client ), display and recording ( see Figure 2 ).
Sub-system 2 : Client and server based GPS coordinates streaming system With the lack of GPS software on the Lenovo 4500U Laptop and the requirement to use a single device to record both the camera video image and GPS coordinates to avoid time synchronization issues , an HTC One X smartphone with a built-in GPS module was used . Using the configured Wi-Fi link between the Lenovo 4500U Laptop and the HTC smartphone , the GPS client programme ( TCP / IP socket ) running inside the smartphone sends the result of GPS coordinate calculation to the connected server ( laptop ). At the same time , the smartphone also displays the current GPS location on a rendered map .
Sub-system 3 : Camera calibration for horizontal ( lane width ) measurement The image produced by a pin-hole camera can be modelled as the projection of an object in the 3D world into a 2D image plane . For example , the projection of a planar object ( e . g . road surface ) in the 3D world into the 2D image plane can be described by a planar homography H ( matrix with dimension 3x3 ) as
[ 1.0 ]
Figure 2 : Sequential diagram ( GPS capture , display and recording ) where m is the pixel coordinate on the camera image plane and M is the grid coordinate of the 3D planar object [ 5 ].
27