Object separation to take ROI (region of interest) is needed on vehicle recognition. There is one layer object separation algorithm to solve this problem but sometime is not enough to separate all objects. This paper contains an advanced model in two layer to separate all objects in binary image. Results showed that the model can improve the border of ROI. The model also probably can be improved into adaptive n-layer separation. The adaptive function is to find when to use horizontal or vertical separation first so it can produce lower n and computing time.

two layer, object separation, binary image.