Method for detecting objects in a warehouse and/or for spatial orientation in a warehouse
A method for detecting objects in a warehouse and/or for spatial orientation in a warehouse includes: acquiring image data with a 3-D camera which is fastened to an industrial truck so that a viewing direction of the 3-D camera has a defined horizontal angle, wherein the 3-D camera has an image sensor with sensor elements arranged matrix-like and the image data comprises a plurality of pixels, wherein distance information is assigned to each pixel, calculating angle information for a plurality of image elements, which each specify an angle between a surface represented by the image element and a vertical reference plane, determining a predominant direction based on the frequency of the calculated angle information, calculating the positions of the of the acquired pixels along the predominant direction, detecting at least one main plane of the warehouse based on a frequency distribution of the calculated positions.
1. A method for detecting objects in a warehouse and/or for spatial orientation in a warehouse having the following steps:
acquiring image data with a 3-D camera which is fastened to an industrial truck so that a viewing direction of the 3-D camera has a defined angle to the horizontal, wherein the 3-D camera has an image sensor with sensor elements arranged matrix-like and the image data comprises a plurality of pixels, wherein distance information is assigned to each pixel,
calculating angle information for a plurality of image elements, which each specify an angle between a surface represented by the image element and a vertical reference plane,
determining a predominant direction based on the frequencies of the calculated angle information,
calculating the positions of the acquired pixels along the predominant direction,
detecting at least one main plane of the warehouse based on a frequency distribution of the calculated positions.
2. The method according to claim 1 , wherein the image elements for which the angle information is calculated, are calculated in each case using averaging over all pixels contained in a predefined volume element.
3. The method according to claim 1 , wherein one of the main planes is a wall plane, which is detected at a first position and perpendicular to the predominant direction, if the frequency distribution at the first position has a first maximum and the number of pixels with positions further remote from the 3-D camera is less than a predefined first limit value.
4. The method according to claim 3 , wherein one of the main planes is a shelf front plane, which is detected at a second position and perpendicular to the predominant direction, if the frequency distribution at the second position has a second maximum and the second position is arranged closer to the 3-D camera by 0.9 m to 1.7 m than the first position.
5. The method according to claim 1 , further comprising one of the main planes is a shelf front plane, which is detected at a third position and perpendicular to the predominant direction, if the frequency distribution has only a single distinct maximum that is arranged at the third position, and the number of pixels with positions further remote from the 3-D camera is greater than a predefined second limit value.
6. The method according to claim 1 , further comprising the further steps:
projecting the pixels, which lie in a surrounding area of a detected main plane onto this main plane
and detecting objects of the warehouse using comparison of the projected, two-dimensional data with predefined templates.
7. The method according to claim 1 , further comprising a vacant warehouse space is detected if no pixels, or respectively segments, are present in a defined spatial region above an object that was detected or classified as a horizontal shelf support.
8. The method according to claim 1 , further comprising the dimensions (b, h, t) of a load are determined by evaluating the pixels, or respectively segments, acquired above an object detected as a pallet.
9. The method according to claim 1 , further comprising the industrial truck has a navigation system that provides position data for the position of the industrial truck in the warehouse, and that the position data are modified considering a detected main plane and/or a detected object.
10. The method according to claim 1 , further comprising the 3-D camera is a time-of-flight camera or a camera system that projects and maps a defined template on the image field and that determines the distance information from a shift of the template.
11. An industrial truck with an apparatus for detecting objects in a warehouse and/or for spatial orientation in a warehouse, wherein the apparatus comprises the following:
a 3-D camera for acquiring image data, wherein the 3-D camera is fastened to the industrial truck so that a viewing direction of the 3-D camera has a defined angle to the horizontal, an image sensor having sensor elements arranged matrix-like and the image data comprises a plurality of pixels, to each of which distance information is assigned,
an evaluation apparatus for evaluating the image data that is designed for the purpose of calculating angle information, which each specify an angle between a surface represented by the image element and a vertical reference plane, for a plurality of image elements, determining a predominant direction based on the frequencies of the calculated angle information, calculating positions of the acquired pixels along the predominant direction, and detecting at least one main plane of the warehouse based on a frequency distribution of the calculated positions.