Abstract: The present invention relates to a computer-implemented method for improving the interpretation of the surroundings of a vehicle (100), wherein the method comprises: receiving a source image (14) of a surrounding environment (118) of the vehicle, which source image is captured by a sensor unit (12a) of the vehicle; receiving or computing a depth data image (20) comprising depth data based on the source image and at least one more source image; detecting a repetitive pattern (30) in the received source image; and based on the detection of the repetitive pattern, determining that an area (32) of the depth data image, which area (32) corresponds to an area (34) of the detected repetitive pattern in the received source image, contains unreliable depth data.

Abstract: A method for improving the interpretation of the surroundings of a UAV, and to a UAV system is presented herein. The method comprises the steps of acquiring an image comprising depth data, and determining a boundary between a first image portion and a second image portion. The second image portion surrounds the first image portion, and the boundary is defined by an interface between first periphery sub-portions and second periphery sub-portions. A difference in depth data between adjacent first periphery sub-portions and second periphery sub-portions is above a first predetermined threshold and/or second periphery sub-portions comprises undefined depth data. The method further comprises determining the area of the first image portion, and if the area of the first image portion is below a second predetermined threshold, determining that the first image portion contains incorrect depth data and deleting and/or adjusting the incorrect depth data contained by the first image portion.

Abstract: The present invention relates to a method for validating sensor units in a UAV. The UAV comprising: a first sensor unit and a second sensor unit, each sensor unit being configured to create an image of the surroundings. The method comprising the steps of: taking a first image by the first sensor unit, taking a second image by the second sensor unit, wherein the second image and the first image at least partly overlap, and comparing the overlapping portions between the first image and the second image. Based on a result in which the overlapping portions of the first image and the second image do not correlate to each other, it is determined that at least one of the first sensor unit and the second sensor unit is dysfunctional.

Abstract: The present invention relates to a computer-implemented method for improving the interpretation of the surroundings of a vehicle (100), wherein the method comprises: receiving a source image (14) of a surrounding environment (118) of the vehicle, which source image is captured by a sensor unit (12a) of the vehicle; receiving or computing a depth data image (20) comprising depth data based on the source image and at least one more source image; detecting a repetitive pattern (30) in the received source image; and based on the detection of the repetitive pattern, determining that an area (32) of the depth data image, which area (32) corresponds to an area (34) of the detected repetitive pattern in the received source image, contains unreliable depth data.

Abstract: A method for improving the interpretation of the surroundings of a UAV, and to a UAV system is presented herein. The method comprises the steps of acquiring an image comprising depth data, and determining a boundary between a first image portion and a second image portion. The second image portion surrounds the first image portion, and the boundary is defined by an interface between first periphery sub-portions and second periphery sub-portions. A difference in depth data between adjacent first periphery sub-portions and second periphery sub-portions is above a first predetermined threshold and/or second periphery sub-portions comprises undefined depth data. The method further comprises determining the area of the first image portion, and if the area of the first image portion is below a second predetermined threshold, determining that the first image portion contains incorrect depth data and deleting and/or adjusting the incorrect depth data contained by the first image portion.