Limits...
A Multispectral Image Creating Method for a New Airborne Four-Camera System with Different Bandpass Filters.

Li H, Zhang A, Hu S - Sensors (Basel) (2015)

Bottom Line: For this multispectral system, an automatic multispectral data composing method was proposed.For the difficult registration problem between visible band images and near-infrared band images in cases lacking manmade objects, we presented an effective method based on the structural characteristics of the system.Experiments show that our method can acquire high quality multispectral images and the band-to-band alignment error of the composed multiple spectral images is less than 2.5 pixels.

View Article: PubMed Central - PubMed

Affiliation: Key Laboratory of 3D Information Acquisition and Application of Ministry, Capital Normal University, Beijing 100048, China. lihanlun@126.com.

ABSTRACT
This paper describes an airborne high resolution four-camera multispectral system which mainly consists of four identical monochrome cameras equipped with four interchangeable bandpass filters. For this multispectral system, an automatic multispectral data composing method was proposed. The homography registration model was chosen, and the scale-invariant feature transform (SIFT) and random sample consensus (RANSAC) were used to generate matching points. For the difficult registration problem between visible band images and near-infrared band images in cases lacking manmade objects, we presented an effective method based on the structural characteristics of the system. Experiments show that our method can acquire high quality multispectral images and the band-to-band alignment error of the composed multiple spectral images is less than 2.5 pixels.

No MeSH data available.


Image registration; (left) the result obtained by using the RANSAC directly; (right) the result obtained by using our method.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4541943&req=5

sensors-15-17453-f009: Image registration; (left) the result obtained by using the RANSAC directly; (right) the result obtained by using our method.

Mentions: As shown in Figure 8, four different band images acquired by the multispectral system at one other exposal are selected. Compared with Figure 4 and Figure 6, Figure 8 contains no artificial objects. The number of SIFT feature points in the four images are 4137, 5578, 5469, and 7385, respectively. For the same reason as in the above two experiments, the green band is chosen as the reference band, and the other three band images are registered to it. As shown in Table 1, the numbers of initial matches (IM) are 555, 792 and 303, and the correct matches (CM) respectively are 342, 525 and 19, thus the correct rate (CR) respectively is 61%, 66% and 6%. The correct rate between the green band and the infrared band is evidently lower than 50%. Therefore the RANSAC cannot be used directly, as shown in Figure 9 (left). Because Figure 4, Figure 6 and Figure 8 are from a same flight strip, they have little change in flying height and attitude. So, the evaluated (xrow, ycol) in Experiment 1 can also be used to eliminate error matches in this experiment. As shown in Table 1, 210, 255 and 277 false matches (RFM) have been eliminated respectively. After that, the number of matches respectively become 345, 537 and 26, and the correct rates (CR) respectively becomes 99%, 97%, 73%; the registration performance is significantly improved. All of these percentages are much higher than 50%, and the RANSAC is more reliable and can be used directly to obtain correct matching results, as shown in Figure 9 (right).


A Multispectral Image Creating Method for a New Airborne Four-Camera System with Different Bandpass Filters.

Li H, Zhang A, Hu S - Sensors (Basel) (2015)

Image registration; (left) the result obtained by using the RANSAC directly; (right) the result obtained by using our method.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4541943&req=5

sensors-15-17453-f009: Image registration; (left) the result obtained by using the RANSAC directly; (right) the result obtained by using our method.
Mentions: As shown in Figure 8, four different band images acquired by the multispectral system at one other exposal are selected. Compared with Figure 4 and Figure 6, Figure 8 contains no artificial objects. The number of SIFT feature points in the four images are 4137, 5578, 5469, and 7385, respectively. For the same reason as in the above two experiments, the green band is chosen as the reference band, and the other three band images are registered to it. As shown in Table 1, the numbers of initial matches (IM) are 555, 792 and 303, and the correct matches (CM) respectively are 342, 525 and 19, thus the correct rate (CR) respectively is 61%, 66% and 6%. The correct rate between the green band and the infrared band is evidently lower than 50%. Therefore the RANSAC cannot be used directly, as shown in Figure 9 (left). Because Figure 4, Figure 6 and Figure 8 are from a same flight strip, they have little change in flying height and attitude. So, the evaluated (xrow, ycol) in Experiment 1 can also be used to eliminate error matches in this experiment. As shown in Table 1, 210, 255 and 277 false matches (RFM) have been eliminated respectively. After that, the number of matches respectively become 345, 537 and 26, and the correct rates (CR) respectively becomes 99%, 97%, 73%; the registration performance is significantly improved. All of these percentages are much higher than 50%, and the RANSAC is more reliable and can be used directly to obtain correct matching results, as shown in Figure 9 (right).

Bottom Line: For this multispectral system, an automatic multispectral data composing method was proposed.For the difficult registration problem between visible band images and near-infrared band images in cases lacking manmade objects, we presented an effective method based on the structural characteristics of the system.Experiments show that our method can acquire high quality multispectral images and the band-to-band alignment error of the composed multiple spectral images is less than 2.5 pixels.

View Article: PubMed Central - PubMed

Affiliation: Key Laboratory of 3D Information Acquisition and Application of Ministry, Capital Normal University, Beijing 100048, China. lihanlun@126.com.

ABSTRACT
This paper describes an airborne high resolution four-camera multispectral system which mainly consists of four identical monochrome cameras equipped with four interchangeable bandpass filters. For this multispectral system, an automatic multispectral data composing method was proposed. The homography registration model was chosen, and the scale-invariant feature transform (SIFT) and random sample consensus (RANSAC) were used to generate matching points. For the difficult registration problem between visible band images and near-infrared band images in cases lacking manmade objects, we presented an effective method based on the structural characteristics of the system. Experiments show that our method can acquire high quality multispectral images and the band-to-band alignment error of the composed multiple spectral images is less than 2.5 pixels.

No MeSH data available.