Recognition of unsegmented targets invariant under transformations of intensity

Abstract

Images taken in noncooperative environments do not always have targets under the same illumination conditions. There is a need for methods to detect targets independently of the illumination. We propose a technique that yields correlation peaks that are invariant under a linear intensity transformation of object intensity. The new locally adaptive contrast-invariant filter accomplishes this by combining three correlations in a nonlinear way. This method is not only intensity invariant but also has good discrimination and resistance to noise. We present simulation results for various intensity transformations with and without random and correlated noise. When the noise is high enough to threaten errors, the method trades off intensity invariance in order to achieve the optimum signal to noise ratio, and the peak to sidelobe ratio in the presence of clutter is always greater than one. In the presence of random disjoint noise, the signal to noise ratio is independent of the target contrast and of the level of the noise.

Asplünd׳s metric defined in the Logarithmic Image Processing (LIP) framework: A new way to perform double-sided image probing for non-linear grayscale pattern matching

“…A recurrent problem of such methods resides in their high sensitivity to illumination variations, pattern size changing and generally small evolutions of image acquisition conditions. In case of variable illumination conditions, D. Lefebvre et al (Lefebvre, 2002) propose a technique of correlation peaks that are invariant under linear intensity transformation.…”

An efficient method for automatic target position detection system

“…This technique is equivalent to projecting the scene onto the subspace orthogonal to ¡(x), that is, to project the zero-mean scene in this region of support and then to calculate the cosine of the angle between the scene [2]. We illustrate the vector interpretation as shown in figure4.…”

“…One illumination model that is widely used has a target that is multiplied by an unknown constant factor so that the correlation peak will change by the same amount; in such cases dark targets can be missed. The two types of intensity transformation: multiplicative and additive are considered for computing the light source intensity changes [1], [2]. Additive-intensity transformation is useful for different camera settings or when scattered light enters the camera.…”

Intensity-invariant nonlinear filtering for detection in camouflage

“…On the other hand, the normalized correlation between a target different from the reference will give a cosine value smaller than one. Lefebvre et al 14 proposed an extension of the previous multiplicative-intensity transformation to include an additive term. Indeed, if the intensity transformation of the targets is af͑x͒ ϩ b, where a and b are unknown parameters that are constant over a single target, the LACIF method is an angular measurement between two vectors that are projected onto the same plane.…”