We propose an improved Lucas-Kanade template tracking method with drift correction, which can be applied in rigid extended object. Due to error accumulation, primary template tracking method leads to template drift and loses object gradually. In order to alleviate template drift, SIFT (Scale Invariant Feature Transform) feature is used to correct the template drift. SIFT feature is invariant to scale, rotation even affine transformation, so, according to matching SIFT key-points between frames, the affine parameters of object transformation can be computed, then the current template position can be obtained by affine parameters and primary template position. The experiment results prove that the improved template tracking method based on SIFT drift correction can more accurately track the rigid extended object and can alleviate the tracking position drifting effectively.
Using the local modeling method can get sub-pixel localization in digital image measurement, in order to improve the
precision of sub-pixel measurement which is usually obtained by computing the properties of fitting curves in local
modeling method, an effective method is presented. The proposed method is that at first using bilinear interpolation or
other more reasonable interpolation ways to original image obtains a new image and then some edge features are
extracted from the new image by chain code edge detection or other ways. Due to using of interpolation, after the pixels
of extracted edge features are mapped to original image size, the edge features will contain more pixels which belong to
integer pixels from original image and the sub-pixels which the interpolation image produces, subsequently those pixels
will be used to fit curve to improve the accuracy of the fitting. According to the result of the improving fitting, a more
accurate measurement is obtained by utilizing the properties of the curves. At the last section of this paper, the proposed
method is evaluated by using real images which is collected by digital camera, the experiment result turns out that the
algorithm owns better accuracy than the one which is only by only fitting on account of the proposed method which
owns better standard deviation than the only fitting way without interpolation.