Moving object detection apparatus and method
Disclosed is directed to a moving object detection apparatus and method. The apparatus comprises an image capture module, an image alignment module, a temporal differencing module, a distance transform module, and a background subtraction module. The image capture module derives a plurality of images in a time series. The image alignment module aligns the images if the image capture module is situated on a movable platform. The temporal differencing module performs temporal differencing on the captured images or the aligned images, and generates a difference image. The distance transform module transforms the difference image into a distance map. The background subtraction module applies the distance map to background subtraction technology and compares the results with the current captured image, so as to obtain the information for moving objects.
1. A moving object detection apparatus, comprising:
an image capture module that captures a plurality of images for one or more moving objects at different times;
an image align module that aligns said plurality of images captured at different times if said image capture module is situated on a movable platform;
a temporal differencing module that performs temporal differencing on said captured or aligned images to obtain a difference image;
a distance transform module that transforms said difference image into a distance map; and
a background subtraction module that applies said distance map to a background subtraction technology and compares with a current captured image to obtain moving object information.
2. The apparatus as claimed in claim 1 , wherein in a movable platform, said image align module provides one or more alignment parameters to said background subtraction module to obtain said moving object information.
3. The apparatus as claimed in claim 1 , wherein in a movable platform, said image align module provides said aligned images to said temporal differencing module.
4. The apparatus as claimed in claim 1 , wherein said moving object information at least includes an area where said one or more moving objects occur in the images.
5. The apparatus as claimed in claim 1 , wherein said distance map provides a probability distribution of a location for said one or more moving objects.
6. The apparatus as claimed in claim 1 , said apparatus is applicable to both movable platform and fixed platform.
7. The apparatus as claimed in claim 6 , wherein in a fixed platform, said apparatus does not comprise an image align module.
8. A moving object detection method performed in a moving object detection apparatus, said apparatus having an image capture module, an image align module, a temporal differencing module, a distance transform module, and a background subtraction module, and said method comprising the steps of:
capturing a plurality of images for one or more moving objects at different times using said image capture module;
aligning said plurality of images captured at different times using said image align module if said captured images are on a movable platform;
performing temporal differencing on said captured or aligned images using said temporal differencing module to obtain a difference image;
transforming said difference image into a distance map using said distance transform module; and
applying said distance map to a background subtraction technology and comparing with a current captured image using said background subtraction module to obtain moving object information.
9. The method as claimed in claim 8 , wherein in a movable platform, at least an alignment parameter is applied to a pair of said captured images for aligning images at different times.
10. The method as claimed in claim 8 , wherein said background subtraction technology at least includes:
applying at least an alignment parameter to a background model for said one or more moving objects for aligning said background model to a current captured image;
updating said current captured image into said aligned background model by using said distance map as an updating rate for the background model updating; and
comparing the aligned background model with said current captured image by taking said distance map as a weighting function to perform foreground detection at current time for detecting said one or more moving objects.
11. The method as claimed in claim 8 , wherein in a movable platform, alignment is performed on neighboring images before temporal differencing to align background area of said neighboring images for compensating image change caused by the movable platform.
12. The method as claimed in claim 8 , wherein said temporal differencing technology is to perform an AND operation on two frame differences obtained from deformation transformation to compute said difference image.
13. The method as claimed in claim 8 , wherein an area where said one or more moving objects occur in an image is seen from said distance map.
14. The method as claimed in claim 10 , wherein in foreground detection, said distance map is used as an adaptive threshold for foreground determination.
15. The method as claimed in claim 10 , wherein in background updating, said distance map is used as an adaptive updating rate.