Unfortunately, in article specified by you this question isn't discussed in that context in which it is formulated.
Statement of a question is connected with dynamic nature of change of temperature of the area observed by means of the IK-camera on a sample which is exposed to periodic influence of sequence of laser impulses. In the traditional formulation the concept NETD is connected with passive registration of a signal from supervision area.
I think it's a bit complicated. First it depends on the linear dynamic range of your camera. If a rise in thermal glow from a temperature rise in the absorbing surface is enough to push the camera into a nonlinear response region, then certainly the NEP / NETD will change. Second, depending on the spectral emissivity of the absorbing surface, a significant rise in its temperature could result in different spectral content of the thermal glow. If the wavelength response of the IR camera is not flat, this could also affect the performance, and the NETD. For a concrete answer, I think you would have to make controlled measurements on the system you are going to study, increasing the absorbing surface temperature.
Unfortunately, I had no opportunity to answer earlier.
Many thanks for your detailed answer. I agree with what you wrote. But I was faced with a more banal reason. When processing the results of measurements, I just do not quite correctly considered the transmission characteristics of the atmosphere during the work in different spectral intervals.