**End**

By applying the SuspiciousEvaluation algorithm, we can identify which segments are suspicious and require further post-processing.

(4) Localized correction

As shown in Figure 2, for a segmen<sup>t</sup> *hi* that is suspected of containing error classified pixels, the post-processing strategy can be described as follows: First, based on *hi.hreci,* create a cut rectangle *cf*, (*cf* = rectangle *hi.hreci* enlarged by β pixels), where β is the number of pixels to enlarge and the default value is 10. Second, use the *cf* cut sub-images from *Iimage* and *Icls* to obtain a *Sub*image and a *Subcls*. Third, input *Sub*image and *Subcls* to the CRF algorithm, and obtain a corrected classification result *Corrected*-*Subcls*. Finally, based on the pixel locations in *hi.hpixelsi,* obtain the pixels from *Corrected*-*Subcls* and write them to *Icls*, which constitutes localized area correction on *Icls*. For the entire *Icls*, the localized correction algorithm on *Icls* is as follows (LocalizedCorrection Algorithm):

**Algorithm** LocalizedCorrection

```
Input: Icls, HList
Output: Icls
Begin
     foreach(hi in HList){
          if (hi.hspci==0) then continue;
          cf= Enlarge rectangle hi.hreci by β pixels;
          Subimage, Subcls = Cut sub-images from Iimage and Icls;
          Corrected-Subcls = Process Subimage and Subcls by CRF algorithm;
          pixels = Obtain pixels in hi.hpixelsi from Corrected-Subcls;
          Icls←pixels;
     }
     return Icls
```