Sparse Representation Classification Based on Flexible Patches Sampling of Superpixels for Hyperspectral Images

Figure 5

Classification maps by different methods for the Salinas image with overall accuracies (OA in %). (a)Three-band color image by PCA. (b) Ground-truth image. (c) Color legend of corresponding classes. (d) PSRC (86.48%). (e) JSRM (92.07%). (f) MJSR (93.64%). (g) MASR (98.06%). (h) BTC-WSL (95.99%). (i) SRC-FPSS (99.35%).

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)