Image Matching Algorithm Based on Exposure and Color Information

Fig. 3. Experimental test images. (a)(b) Book matching images; (c)(d) square images with different angles; (e)(f) scenery building images with different angles

Download full size

Fig. 4. Image matching results of Harris-SIFT algorithm with exposure and color information. (a) Matching of Figs. 3 (a) and (b); (b) matching of Figs. 3(c) and (d); (c) matching of Figs. 3(e) and (f)

Download full size

Fig. 5. Image matching results for Harris-SIFT algorithm. (a) Matching of Figs. 3 (a) and (b); (b) matching of Figs. 3(c) and (d); (c) matching of Figs. 3(e) and (f)

Download full size

Table 1. Statistic of feature point detection time of two algorithms
View table
Table 1. Statistic of feature point detection time of two algorithms
Figure Number of feature points Time of feature points detection /s
Harris-SIFT Proposed algorithm Harris-SIFT Proposed algorithm
Fig.3 (a) 643 987 3.31 2.94
Fig.3 (c) 315 526 1.49 1.37
Fig.3 (e) 884 1245 3.47 2.96

Table 2. Accuracy and cost time of the proposed algorithm

View table

Table 2. Accuracy and cost time of the proposed algorithm

Figure	Number offeature points	Number ofmatches	Number ofcorrect matches	Percent of correctmatches /%	Totaltime /s
Fig.3 (a)	987	232	214	92.24	4.53
Fig.3 (b)		232	214	92.24	1186
Fig.3 (c)	526	103	92	89.32	1.56
Fig.3 (d)		103	92	89.32	476
Fig.3 (e)	1245	146	129	88.36	4.09
Fig.3 (f)		146	129	88.36	1023