Fig. 1. Flow chart of proposed road extraction method from remote sensing image

Fig. 2. Shape analysis of elliptic structural elements. (a) λ1≈λ2≈0; (b) λ1≫λ2≈0; (c) λ1≈λ2≫0

Fig. 3. Elliptic structural elements

Fig. 4. Adaptive morphological top and bottom hat transformation. (a) Grayscale image; (b) result of fixed structural elements; (c) result of adaptive morphological structural elements

Fig. 5. Improved aspect ratio. (a) Road target area; (b) new rectangle

Fig. 6. Post processing of road segmentation. (a) Binary image; (b) shape feature extraction; (c) morphological filtering; (d) overlay effect image

Fig. 7. Road extraction simulation experiment. (a) Original image(from left to right: I₁、I₂、I₃、I₄); (b) circular structural element; (c) linear structural element; (d) watershed algorithm of morphological gradient reconstruction; (e) proposed method; (f) effect of road superimposed with original images; (g) extraction results of manual images

Table 1. Error comparison of road extraction