Flatness detecting method of road, computing apparatus, and computer-readable medium
A flatness detecting method of road, a computing apparatus, and a computer- readable medium are provided. In the method, depth information is obtained. The depth information includes depth values corresponding to a plurality of pixels. The depth information is converted into a height relation map. Height values of the pixels in the height relation map are converted from the corresponding depth values, and the height value of each of the pixels is related to a height of a ground. According to the height relation map, a flatness category corresponding to the pixels is determined via a semantic segmentation model. The semantic segmentation model is trained with an association 10 between one or a plurality of reference heights and corresponding flatness categories. Accordingly, a road condition may be effectively detected.
1 . A flatness detecting method of road, comprising:
obtaining depth information, wherein the depth information comprises depth values corresponding to a plurality of pixels;
converting the depth information into a height relation map, wherein a height value of the pixels in the height relation map is converted from a corresponding depth value, and the height value of each of the pixels is related to a height of a ground;
determining flatness categories corresponding to the pixels via a semantic segmentation model according to the height relation map, wherein the semantic segmentation model is trained with an association between at least one reference height and a corresponding flatness category,
determining that a flatness category of a region to be measured is slope up; and
determining a slope up angle of the region to be measured via a first function corresponding to the slope up, wherein the first function is:
α
=
tan
-
1
(
HA
2
-
HA
1
)
(
DIS
2
-
DIS
1
)
,
α is the slope up angle, HA1 is a first reference height, HA2 is a second reference height, DIS1 is a first distance corresponding to the first reference height, and DIS2 is a second distance corresponding to the second reference height.
2 . The flatness detecting method of road of claim 1 , wherein the step of converting the depth information into the height relation map comprises:
determining a projection height of a depth value of a first pixel in the pixels at a projection angle, wherein the projection angle is obtained based on a position of the first pixel in the pixels; and
determining the height value of the first pixel in the height relation map according to a difference between a sensing height and the projection height, wherein the sensing height is a height relative to the ground when the depth information is obtained.
3 . The flatness detecting method of road of claim 2 , further comprising:
defining a depth value of the first pixel as a length of a first side of a triangle;
defining the projection height as a length of a second side of the triangle; and
defining the projection angle as an included angle between the first side and the second side.
4 . The flatness detecting method of road of claim 3 , further comprising:
determine a distance of a region to be measured corresponding to the first pixel according to the projection angle and the first side.
5 . The flatness detecting method of road of claim 2 , further comprising:
obtaining reference depth information, wherein the reference depth information comprises depth values of the pixels when the ground is flat;
matching the reference depth information with a plurality of reference angles to obtain a matching result, wherein each of the pixels in the matching result matches a reference angle; and
determining the projection angle corresponding to the first pixel according to the matching result.
6 . The flatness detecting method of road of claim 1 , wherein the flatness category predicted by the semantic segmentation model comprises a protrusion, a depression, the slope up, and a slope down.
7 . The flatness detecting method of road of claim 1 , wherein the step of determining the flatness categories corresponding to the pixels via the semantic segmentation model comprises:
dividing the pixels into a plurality of regions; and
determining flatness categories corresponding to the regions of the pixels respectively via the semantic segmentation model.
8 . The flatness detecting method of road of claim 1 , further comprising:
determining that a flatness category of a region to be measured is a slope down; and
determining a slope down angle of the region to be measured via a second function corresponding to the slope down, wherein the second function is:
β
=
tan
-
1
(
HA
4
-
HA
3
)
(
DIS
4
-
DIS
3
)
,
β is the slope down angle, HA3 is a third reference height, HA4 is a fourth reference height, DIS3 is a third distance corresponding to the third reference height, and DIS4 is a fourth distance corresponding to the fourth reference height.
9 . The flatness detecting method of road of claim 1 , further comprising:
generating a control command for controlling a mobile vehicle according to the height relation map and a corresponding flatness category thereof, wherein the control command is related to at least one of a travel speed, a rotation, and a stop of the mobile vehicle.
10 . A computing apparatus, comprising:
a memory storing a program code; and
a processor coupled to the memory and loading the program code and executing:
obtaining depth information, wherein the depth information comprises depth values corresponding to a plurality of pixels;
converting the depth information into a height relation map, wherein height values of the pixels in the height relation map are converted from corresponding depth values, and the height value of each of the pixels is related to a height of a ground;
determining flatness categories corresponding to the pixels via a semantic segmentation model according to the height relation map, wherein the semantic segmentation model is trained with an association between at least one reference height and a corresponding flatness category,
determining that a flatness category of a region to be measured is a slope up; and
determining a slope up angle of the region to be measured via a first function corresponding to the slope up, wherein the first function is:
α
=
tan
-
1
(
HA
2
-
HA
1
)
(
DIS
2
-
DIS
1
)
,
α is the slope up angle, HA1 is a first reference height, HA2 is a second reference height, DIS1 is a first distance corresponding to the first reference height, and DIS2 is a second distance corresponding to the second reference height.
11 . The computing apparatus of claim 10 , wherein the processor further executes:
determining a projection height of a depth value of a first pixel in the pixels at a projection angle, wherein the projection angle is obtained based on a position of the first pixel in the pixels; and
determining the height value of the first pixel in the height relation map according to a difference between a sensing height and the projection height, wherein the sensing height is a height of a depth camera obtaining the depth information relative to the ground.
12 . The computing apparatus of claim 11 , wherein the processor further executes:
defining the depth value of the first pixel as a length of a first side of a triangle;
defining the projection height as a length of a second side of the triangle; and
defining the projection angle as an included angle between the first side and the second side.
13 . The computing apparatus of claim 12 , wherein the processor further executes:
determining a distance of a region to be measured corresponding to the first pixel according to the projection angle and the first side.
14 . The computing apparatus of claim 11 , wherein the processor further executes:
obtaining reference depth information, wherein the reference depth information comprises depth values of the pixels when the ground is flat;
matching the reference depth information with a plurality of reference angles to obtain a matching result, wherein each of the pixels in the matching result matches a reference angle; and
determining the projection angle corresponding to the first pixel according to the matching result.
15 . The computing apparatus of claim 10 , wherein the flatness category predicted by the semantic segmentation model comprises a protrusion, a depression, the slope up, and a slope down.
16 . The computing apparatus of claim 10 , wherein the processor further executes:
dividing the pixels into a plurality of regions; and
determining flatness categories corresponding to the regions of the pixels respectively via the semantic segmentation model.
17 . The computing apparatus of claim 10 , wherein the processor further executes:
determining that a flatness category of a region to be measured is a slope down; and
determining a slope down angle of the region to be measured via a second function corresponding to the slope down, wherein the second function is:
β
=
tan
-
1
(
HA
4
-
HA
3
)
(
DIS
4
-
DIS
3
)
,
β is the slope down angle, HA3 is a third reference height, HA4 is a fourth reference height, DIS3 is a third distance corresponding to the third reference height, and DIS4 is a fourth distance corresponding to the fourth reference height.
18 . A non-transitory computer-readable medium, loaded with a program code via a processor to perform the following steps:
obtaining depth information, wherein the depth information comprises depth values corresponding to a plurality of pixels;
converting the depth information into a height relation map, wherein height values of the pixels in the height relation map are converted from corresponding depth values, and the height value of each of the pixels is related to a height of a ground;
determining flatness categories corresponding to the pixels via a semantic segmentation model according to the height relation map, wherein the semantic segmentation model is trained with an association between at least one reference height and a corresponding flatness category,
determining that a flatness category of a region to be measured is slope up; and
determining a slope up angle of the region to be measured via a first function corresponding to the slope up, wherein the first function is:
α
=
tan
-
1
(
HA
2
-
HA
1
)
(
DIS
2
-
DIS
1
)
,
β is the slope up angle, HA1 is a first reference height, HA2 is a second reference height, DIS1 is a first distance corresponding to the first reference height, and DIS2 is a second distance corresponding to the second reference height.