Fig. 1. Method for determining best image definition

Fig. 2. Design of test platform for mobile robot visual navigation

Fig. 3. Image definition characteristics under different CCD optical integration time. (a) 5 ms; (b) 15 ms; (c) 50 ms; (d) 90 ms

Fig. 4. Image definition characteristics under different CCD defocus statuses. (a) Defocused status 1; (b) defocused status 2; (c) defocused status 3; (d) defocused status 4

Fig. 5. Response curves of best image definition under different CCD optical integration time. (a) TenenGrad; (b) average gradient; (c) information entropy

Fig. 6. Response curves of best image definition under different CCD defocus statuses. (a) TenenGrad; (b) average gradient; (c) information entropy

Fig. 7. Comparison of response results of definition calculation under different image definition evaluation functions. (a) Sequence of image definition change; (b) response results of definition evaluation functions

Fig. 8. Fuzzy connectedness between two points of a and b in α-adjacent affinity function

Fig. 9. Delineation method of target and background area based on relative fuzzy connectedness

Fig. 10. Reclassification method of pixels in undefined area of images in F-R model

Fig. 11. Flow chart of ROI detection based on fuzzy-rough set model

Fig. 12. ROI segmentation and comparison experiment of image ROI. (a) CCD high definition image; (b) original gray image; (c) result by K-means clustering algorithm(k=2); (d) result by watershed algorithm

Fig. 13. Navigation path recognition result based on F-R model. (a) Seed definition area; (b) final ROI

Fig. 14. Comparison of navigation path recognition results based on F-R model

Fig. 15. Path navigation trajectory of mobile robot. (a) Test paths; (b) robot walking control

Fig. 16. Path navigation trajectory of mobile robot.(a) Path 1; (b) path 2