SAR image speckle suppression method based on DCA algorithm

Abstract

The invention provides an SAR image speckle suppression method based on a DCA algorithm. The SAR image speckle suppression method comprises the steps that a novel second-order operator is constructed; according to the novel second-order operator, an SAR image speckle suppression non-convex variation model with an adaptive regular term is constructed, and the model comprises a novel second-order operator, a first-order term and a data item; the non-convex variation model is transformed, and solving is performed by adopting a DCA algorithm to obtain a solved non-convex variation model; and an SAR image with speckle is input, and a result after speckle noise suppression is obtained by using the solved non-convex variation model. According to the SAR image speckle suppression method with the self-adaptive regular term based on the DCA algorithm, a speckle suppression result with good feature maintenance and smooth region recovery can be obtained, and the method has very good applicability, effectiveness, stability and expansibility.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a method for suppressing SAR image coherent specks based on a DCA algorithm. Background technique [0002] SAR images are detected by the sensors mounted on mobile platforms such as aircrafts and satellites to actively emit microwave signals from targets, so as to receive echo signals with certain technical means for imaging. The SAR system can transmit microwave signals all-weather and all-day, so as to avoid the influence of many external conditions such as night and bad weather. Therefore, the application scenarios of SAR images are very wide, especially for some tasks that require timeliness and long-term regular observation. For example: military reconnaissance, emergency rescue and disaster relief, surface scientific research, etc. With the improvement of SAR system imaging technology, the spatial resolution of the obtained SAR images has also been greatly improved, and th...

AI Technical Summary

Problems solved by technology

But it will inevitably introduce some artificial phenomena, such as staircase phenomenon (TV), fuzzy edge (TGV), etc.

Data-driven methods, they can suppress coherent speckle noise very well, but the results will be very dependent on the selected training data set

As shown above, the excellent coherent speckle noise suppression methods mentioned above will have some artifacts, such as ghosting, brush and boundary blur, etc.

Images