Author Affiliations
College of Electrical and Electronic Engineering, Shandong University of Technology, Zibo, Shandong 255000, Chinashow less
Fig. 1. Processing of transmitted images under LIP model
Fig. 2. Under the LIP model, two images are added together
Fig. 3. Processing process of φ function
Fig. 4. Processing process of φ-1 function
Fig. 5. Four-direction Sobel operator templates. (a) x-axis direction template; (b) y-axis direction template; (c) 45° direction template; (d) 135° direction template
Fig. 6. 3 pixel×3 pixel area
Fig. 7. Gradient image comparison results. (a) Test image; (b) gradient image of proposed algorithm; (c) gradient image of Ref. [14] algorithm; (d) traditional Sobel algorithm gradient image
Fig. 8. Image of real experimental platform
Fig. 9. Internal structure of the experimental platform
Fig. 10. Principle block diagram of experimental platform
Fig. 11. Image of measured target. (a) Car image; (b) triangle base image; (c) triangle base and number plate combination image
Fig. 12. Results of algorithm flow. (a) Original image; (b) filtered image; (c) gradient image; (d) final result image
Fig. 13. Comparison of edge detection results. (a) Original image; (b) traditional Canny algorithm; (c) Sobel algorithm; (d) Ref. [14] algorithm; (e) proposed algorithm
Picture name | Algorithm |
---|
Traditional Canny | Sobel | Ref. [14] | Proposed algorithm |
---|
Car | 1.148 | 0.438 | 1.834 | 2.683 | Triangle base | 0.876 | 0.334 | 1.423 | 2.376 | Base and number plate | 0.852 | 0.339 | 1.372 | 2.215 |
|
Table 1. Comparison of calculation time unit: s
Algorithm | Le | 1-N | 1-F | Me |
---|
Traditional Canny | 0.845 | 0.694 | 0.834 | 0.781 | Sobel | 0.716 | 0.741 | 0.723 | 0.728 | Ref. [14] algorithm | 0.894 | 0.872 | 0.886 | 0.883 | Proposed algorithm | 0.932 | 0.905 | 0.912 | 0.915 |
|
Table 2. Evaluation of algorithm’s performance