Visual robot navigation in outdoor environments would benefit from an illumination-independent representation of images. We explore how such a representation, comprising a black skyline of objects in front of a white sky, can be obtained from dual-channel spectral contrast measures. Light from sky and natural objects under different conditions of illumination was analyzed by five spectral channels: ultraviolet, blue, green, red, and near infrared. Linear discriminant analysis was applied to determine the optimal linear separation between sky and object points. A statistical comparison shows that contrasts with large differences in the wavelength of the two channels, specifically ultraviolet-infrared, blue-infrared, and ultraviolet-red, yield the best separation. Within a single channel, the best separation was obtained for ultraviolet light. The gain in separation quality when all five channels were included is relatively small.
Shuwen Wei, Michael Kam, Yaning Wang, Justin D. Opfermann, Hamed Saeidi, Michael H. Hsieh, Axel Krieger, and Jin U. Kang J. Opt. Soc. Am. A 39(4) 655-661 (2022)
You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
You do not have subscription access to this journal. Figure files are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
You do not have subscription access to this journal. Article tables are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
You do not have subscription access to this journal. Equations are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
Filter Combination, Sensitivity Peak, and Bandwidth (50% to 50% Sensitivity) for All Five Sensor Channels UV, B, G, R, and IRa
Channel
Filter Type (Schott) and Thickness
Peak
Bandwidth
UV
UG11 , BG40
B
BG12 , GG400 , BG28
G
BG7 , GG475
R
RG645 , KG5 , RG665
IR
RG780 , KG4
The sensitivity curves are obtained from the combined filter transmittance (measured data) and the spectral sensitivity of the photodiodes (catalog data).
Table 2
Discriminant Criteria J from Eq. (6) and from Eq. (8), with and the Frequency of Misclassified Data Points for Different Contrastsa
UV
B
G
R
IR
All
J
4.36
3.35
2.37
0.67
0.11
8.99
4.26
3.27
2.32
0.66
0.11
6.95
34
47
76
248
379
15
J
B
G
R
IR
B
G
R
IR
B
G
R
IR
UV
5.15
5.54
6.39
7.30
UV
4.51
4.97
5.92
6.82
UV
44
42
28
15
B
5.37
6.03
6.67
B
3.90
5.34
6.09
B
36
17
8
G
5.38
5.97
G
4.34
5.25
G
21
15
R
5.26
R
2.68
R
55
Top, single-channel contrasts and five-channel contrast (“All”). Bottom, dual-channel contrasts.
Filter Combination, Sensitivity Peak, and Bandwidth (50% to 50% Sensitivity) for All Five Sensor Channels UV, B, G, R, and IRa
Channel
Filter Type (Schott) and Thickness
Peak
Bandwidth
UV
UG11 , BG40
B
BG12 , GG400 , BG28
G
BG7 , GG475
R
RG645 , KG5 , RG665
IR
RG780 , KG4
The sensitivity curves are obtained from the combined filter transmittance (measured data) and the spectral sensitivity of the photodiodes (catalog data).
Table 2
Discriminant Criteria J from Eq. (6) and from Eq. (8), with and the Frequency of Misclassified Data Points for Different Contrastsa
UV
B
G
R
IR
All
J
4.36
3.35
2.37
0.67
0.11
8.99
4.26
3.27
2.32
0.66
0.11
6.95
34
47
76
248
379
15
J
B
G
R
IR
B
G
R
IR
B
G
R
IR
UV
5.15
5.54
6.39
7.30
UV
4.51
4.97
5.92
6.82
UV
44
42
28
15
B
5.37
6.03
6.67
B
3.90
5.34
6.09
B
36
17
8
G
5.38
5.97
G
4.34
5.25
G
21
15
R
5.26
R
2.68
R
55
Top, single-channel contrasts and five-channel contrast (“All”). Bottom, dual-channel contrasts.