Fig. 1. Flow chart of proposed algorithm

Fig. 2. Ideal edge step model. (a) Original image edge; (b) image edge after rotation

Fig. 3. Edge contour discrimination algorithm flow

Fig. 4. Block diagram of improved Zernike moment algorithm

Fig. 5. Schematic diagram of MBR algorithm seeking point set

Fig. 6. Bone stick minimum external moment results. (a) Contour circumscribed rectangle; (b) minimum circumscribed moment of contour; (c) minimum circumscribed moment of bone stick image

Fig. 7. Schematic diagram of the imaging model

Fig. 8. Detection results of different edge detection algorithms. (a) Original drawings of bone stick; (b) Sobel algorithm; (c) Canny algorithm; (d) interpolation algorithm; (e) Zernike algorithm; (f) STZ algorithm

Fig. 9. Local rendering of edge of bone stick. (a) Interpolation algorithm; (b) Zernike algorithm; (c) STZ algorithm

Fig. 10. Bone stick image coordinate point error. (a) X-direction error; (b) Y-direction error

Fig. 11. Bone stick size measurement. (a) Actual length and measured length; (b) actual width and measured width

Fig. 12. Absolute error of measurement size. (a) Length direction; (b) width direction

Fig. 13. Mean absolute error of bone stick length and width measurement

Fig. 14. Relative error of measurement size. (a) Length direction; (b) width direction

Fig. 15. Average relative error of bone stick length and width measurement

Fig. 16. Root mean square error of bone stick length and width measurement