Method and system to characterize and monitor the sharpness of a digital imaging system

A method to determine the sharpness of a digital imaging system, comprising the steps of:

• receiving at least one digital image that is acquired on said digital imaging system as input, said digital image comprising transfer function modulated quantum-noise,
• selecting from said digital images at least one selected image I[X,Y] that comprises a contiguous and substantially equally exposed image area,
• determining at least one ROI in said selected image I[X,Y] that comprises a contiguous and homogeneously exposed image area,
• assigning said ROI as a spatial sample area R[x′,y′] in a calculation method to determine a point spread function PSFxy of said digital image detector system, comprising the steps of:
• acquiring a set of image pixel values comprising transfer function modulated quantum noise in a spatial sample area R[x′,y′] of a digital image I[X,Y] that is substantially equally exposed,
• determining a peripheral sensing kernel ki×j wherein the pixel values matrix k[x,y] is centred on a central pixel,
• calculating for each local image pixel L(x,y) in said sample area R[x′,y′]:
• a pixel deviation ΔL between a pixel value at pixel L(x,y) and its background value BL, wherein BL is defined as the median or mean value of all values in a background region centred on said local pixel L(x,y), and
• a median or mean pixel deviation ΔNd of all background normalized neighbouring pixels values Nd(1→n)=Nd1, Nd2, . . . , Ndn that are located at the same distance d from said local image pixel L(x,y) and that are normalized against their respective background value BNd(i), wherein BNd(i) is defined as the median value of all pixel values in a background region BNd(i) centred on said neighbouring pixels Nd(i),
• calculating an autocorrelation peak value by autocorrelating ΔL for all local pixels L(x,y) in the sample area R[x′,y′], as
• AutoCorrelationPeak(R[x′,y′])=ΣΣx=0,y=0R[x′,y′]ΔL2
• calculating PSFxy as
• PS
• ⁢
• F
• ⁡
• (
• N
• d
• )
• =
• Correlation
• ⁢
• ⁢
• (
• Δ
• ⁢
• ⁢
• L
• ,
• Δ
• ⁢
• N
• _
• d
• )
• Auto
• ⁢
• ⁢
• CorrelationPeak
• ⁢
• ⁢
• (
• R
• ⁡
• [
• x
• ′
• ,
• y
• ′
• ]
• )
• for all values of d from a central pixel as defined in peripheral sensing kernel ki×j, where Correlation(ΔL,ΔNd) is obtained by correlating ΔL with ΔNd for a distance d from local pixel L(x,y), and by summing the results over all pixels in the sample area R[x′,y′],
• Correlation(ΔL,ΔNd)=ΣΣx=0,y=0R[x′,y′]ΔL·ΔNd.