Apparatus, method and computer program for computer vision

1. An apparatus comprising circuitry configured to:

receive motion information regarding a scene obtained from a plurality of motion sensors, the plurality of motion sensors including a first n notion sensor and a second motion sensor, the first and second motion sensors being at least two of a first type of motion sensor or at least one of the first type of motion sensor and at least one of a second type of motion sensor, the second type of motion sensor detecting a different parameter of the scene than the first type of motion sensor;

transform parameters of an error model into an auxiliary representation P, b according to:

P=Σ −1 and

b=Σ −1 μ=Pμ

where μ as a mean value and Σ is a covariance matrix;

apply the error model to data from the first motion sensor to obtain a first motion map and to data from the second motion sensor to obtain a second motion map; and

fuse the first motion map and the second motion map into a fused motion map.

2. The apparatus of claim 1 , wherein the circuitry is further configured to align data of the first motion sensor and align data of the second motion sensor and assign aligned data to cells of the fused motion map.

3. The apparatus of claim 1 wherein the the first motion sensor is a Doppler radar and the second motion sensor is a stereo camera.

4. The apparatus of claim 3 , wherein motion information obtained from the Doppler radar comprises polar coordinates of a cell in the polar coordinate space and angular and radial components of the velocity attributed to the cell.

5. The apparatus of claim 1 , wherein the at least one of the first and second motion sensors is a Doppler radar.

6. The apparatus of claim 1 , wherein at least one of the first and second motion sensors is a stereo camera and the circuitry is configured to obtain scene flow estimation on images captured by the stereo camera.

7. The apparatus of claim 6 , wherein motion information obtained from scene flow estimation comprises image positions, disparity data and displacement data.

8. The apparatus of claim 1 , wherein the circuitry is configured to reconstruct position and velocity from sensor data to obtain first and second motion maps.

9. The apparatus of claim 1 wherein the circuitry is configured to represent the motion map represented by μ v x ,v z , Σ v x ,v z ; x, z are 2d Cartesian coordinates in the motion map, are the mean values of a normal distribution describing a velocity estimate (v x , v z ), and Σ v x , v z is the covariance matrix that represents the information on the motion measurement obtained from a sensor-specific error model.

10. The apparatus of claim 1 , wherein the circuitry is configured to fuse first and second motion mans cell-wise by determining for each cell a joint probability for the velocity.

11. The apparatus of claim 1 wherein the circuitry is configured to assume, for every grid cell of the first and second motion maps, that the velocity v of contacts in this cell follow a normal distribution.

12. The apparatus of claim 1 , wherein the circuitry is configured to fuse information from the first motion map of the first motion sensor “a” and the second motion map of the second motion sensor “b” as

p ( v|a,b )= N (μ c ,Σ c )∝ p ( v|a ) p ( v|b )= N (μ a ,Σ a ) N (μ b ,Σ b )

with

p ( v|a )= N (μ a ,Σ a ) and p ( v|b )= N (μ b ,Σ b )

being two given distributions for v in the same cell and where μ a , μ b are the respective mean values of the normal distribution and Σ a , Σ b are the respective covariance matrices of the normal distribution N.

13. Vehicle control system comprising the apparatus of claim 1 .

14. Vehicle control system of claim 13 wherein the vehicle is a motor vehicle, an electric vehicle, a hybrid vehicle, a robot, an autonomous robot, a drone product, or an autonomous drone product.

15. Advanced driver assistance system comprising the apparatus of claim 1 .

16. A method comprising:

receiving motion information regarding a scene obtained from a plurality of motion sensors, the plurality of motion sensors including a first motion sensor and a second motion sensor, the first and second motion sensors being at least two of a first type of motion sensor or at least one of the first type of motion sensor and at least one of a second type of motion sensor, the second type of motion sensor detecting a different parameter of the scene than the first type of motion sensor; and

fusing information from the first motion sensor “a” and the second motion sensor “b” as

p ( v|a,b )= N (μ c ,Σ c )∝ p ( v|a ) p ( v|b )= N (μ a ,Σ a ) N (μ b ,Σ b )

with

p ( v|a )= N (μ a ,Σ a ) and p ( v|b )= N (μ b ,Σ b )

being two given distributions for v in the same cell and where μ a , μ b are the respective mean values of the normal distribution and Σ a , Σ b are the respective covariance matrices of the normal distribution N into a fused motion map.

17. A non-transitory computer readable storage device having computer readable instructions which, when carried out on a processor, cause the processor to:

receive motion information regarding a scene obtained from a plurality of motion sensors, the plurality of motion sensors including a first motion sensor and a second motion sensor, the first and second motion sensors being at least two of a first type of motion sensor or at least one of the first type of motion sensor and at least one of a second type of motion sensor, the second type of motion sensor detecting a different parameter of the scene than the first type of motion sensor; and

fuse information from the first emotion sensor “a” and the second motion sensor “b” as

p ( v|a,b )= N (μ c ,Σ c )∝ p ( v|a ) p ( v|b )= N (μ a ,Σ a ) N (μ b ,Σ b )

with

p ( v|a )= N (μ a ,Σ a ) and p ( v|b )= N (μ b ,Σ b )

being two given distributions for v in the same cell and where μ a , μ b are the respective mean values of the normal distribution and Σ a , Σ b are the respective covariance matrices of the normal distribution N into a fused motion map.

18. Vehicle control system comprising the non-transitory computer readable storage device of claim 17 .