Disclosure of Invention
The technical problem to be solved by the invention is to provide a warehouse logistics management method aiming at the defects in the prior art, the method has simple steps, reasonable design and convenient realization, can be effectively applied to warehouse logistics, realizes visual management and dynamic management, improves the management efficiency, has obvious effect and is convenient to popularize.
In order to solve the technical problems, the invention adopts the technical scheme that: a warehouse logistics management method, comprising the steps of:
firstly, carrying out route design on an unmanned aerial vehicle according to a storage range and aerial photography parameters;
secondly, the unmanned aerial vehicle acquires aerial images according to the air route;
step three, carrying out image preprocessing on the collected image;
generating a three-dimensional model according to the preprocessed image;
generating DEM data and DOM data according to the three-dimensional model;
step six, forming a real-scene three-dimensional map of the storage area through data fusion of the three-dimensional model, the DEM data and the DOM data;
and seventhly, visually managing the warehouse logistics through the live-action three-dimensional map.
In the warehouse logistics management method, in the first step, the warehouse range includes a whole three-dimensional geographic space for goods storage and flow, and the aerial photography parameters include ground image resolution, image overlap, altitude maintenance, photo inclination and photo rotation angle.
In the warehouse logistics management method, in the second step, the unmanned aerial vehicle acquires aerial images according to air routes and adopts aerial triangulation, and the aerial triangulation comprises encryption point sprint selection, encryption point observation and calculation and encryption process control.
In the above warehouse logistics management method, the encrypted point observation and calculation includes a relative directional tolerance and a model connection tolerance, the relative directional tolerance includes a standard point up-down parallax and a check point up-down parallax, the standard point up-down parallax is less than 0.003mm, and the check point up-down parallax is less than 0.007 mm; the limited difference of the model connection difference comprises poor plane position and poor elevation, and the poor plane position delta S is 0.5F
sm-3d, said elevation being poor
Wherein, Δ S is the poor plane position, Δ H is the poor elevation, F
smIs aerial photography scale denominator, d is photo baseline, f
lIs the focal length of the aerial camera.
In the aforementioned warehouse logistics management method, the image preprocessing on the collected image in step three includes image enhancement on the low-illumination image and image stitching on the enhanced image, and the specific process of image enhancement includes:
step 3011, estimating an illumination component of the low-level image by using gaussian weighted bilateral filtering;
3012, performing adaptive nonlinear stretching on the saturation component of the low-light image in the HSV color space to obtain an enhanced saturation component;
3013, performing single-scale Retinex algorithm enhancement on the brightness component of the low-light image in the HSV color space to obtain a reflection component;
step 3014, taking an index of the reflection component to obtain an enhanced brightness component;
3015, combine the hue with the enhanced saturation component and the enhanced brightness component, and convert the HSV color space into an RGB space;
and 3016, performing contrast correction by using a global adaptive logarithm enhancement algorithm to obtain a final enhanced image.
In the warehouse logistics management method, the image stitching process includes:
step 3021, processing the enhanced image by using an image matching algorithm;
step 3022, extracting homonymous feature points contained in an overlapping area in adjacent images in the same aerial photography zone and homonymous feature points contained in an overlapping area in adjacent images in the adjacent aerial photography zone in the sequence image;
step 3023, performing Euclidean reconstruction on the obtained homonymous feature points and performing triangularization in a three-dimensional space;
and step 3024, generating point cloud data of the warehoused goods through the sequence images of the warehoused goods.
In the warehouse logistics management method, the specific process of generating the three-dimensional model according to the preprocessed image in the fourth step includes:
step 401, point cloud calculation and registration
Solving transformation parameters among all frames of images, superposing and matching multi-frame images acquired at different time, angles and illumination intensities into a unified coordinate system by taking a common part of a storage scene as a reference for image registration, calculating corresponding translation vectors and rotation matrixes, and eliminating redundant information;
step 402, data fusion
Dividing the point cloud space into a plurality of small cubes through a volume grid;
step 403, surface Generation
And (3) generating a three-dimensional model surface by using a real warehousing scene picture acquired by the oblique image as a three-dimensional model map and texture mapping.
In the warehouse logistics management method, the concrete process of generating DEM data according to the three-dimensional model in the fifth step includes: and extracting characteristic points, lines and surface network construction interpolation according to the three-dimensional point cloud to generate the DEM.
In the fifth step, the specific process of generating DOM data according to the three-dimensional model includes:
step 501, introducing a space-three encryption result to establish a test area file and restore a model;
step 502, defining an operation area of a single model, generating an epipolar line image, and matching the epipolar line image to form a matching point and an equal parallax curve;
step 503, checking the matching result;
step 504, according to the elements of the inner and outer directions of the image and the resolution of the image, correcting and resampling the image by adopting a differential correction method to generate a single-model DOM;
step 505, adjusting hue and color;
step 506, setting an embedding range according to the outline coordinates, appointing a storage path, executing an image embedding command, and splicing into a whole DOM.
The invention also discloses a warehouse logistics management system, which comprises:
the unmanned aerial vehicle with the camera is used for aerial photography and acquiring a collected image;
a memory for storing a computer program for implementing a warehouse workflow management method;
a processor for executing the computer program in the memory;
and the display is used for displaying the real-scene three-dimensional map of the storage area obtained by the processor.
Compared with the prior art, the invention has the following advantages:
1. the method has simple steps, reasonable design and convenient realization.
2. The intelligent management system adopts the unmanned aerial vehicle to carry out intelligent management on the warehouse logistics, reduces the working intensity of managers and improves the management efficiency.
3. According to the invention, the real-scene three-dimensional map of the storage area is formed by aerial photography of the unmanned aerial vehicle, image preprocessing and data fusion of the three-dimensional model, DEM data and DOM data, so that visual management of storage logistics is realized, target goods can be accurately and quickly found, and the stock measurement and calculation can be carried out on various goods.
4. According to the image enhancement method in the image preprocessing, the bilateral filtering is adopted to estimate the irradiation component, so that the halo phenomenon generated by the traditional algorithm is effectively avoided, and the edge blurring is prevented; according to the problem that the contrast of the image is not sufficiently improved, the characteristics of a logarithmic function are utilized, the global adaptive logarithm enhancement algorithm is adopted for correction, the improvement is carried out without influencing the high-illumination image, the estimation error of the illumination intensity is effectively reduced, the generation of halation is avoided, the improvement of the contrast plays a role in obviously enhancing the dark area of the low-illumination image, and a good foundation is laid for the subsequent image analysis.
5. The invention can be effectively applied to warehouse logistics, realizes visual management and dynamic management, improves the management efficiency, has obvious effect and is convenient to popularize.
In conclusion, the method has the advantages of simple steps, reasonable design and convenient implementation, can be effectively applied to warehouse logistics, is combined with a management system to realize visual management and dynamic management, improves the management efficiency, has obvious effect and is convenient to popularize.
Detailed Description
As shown in fig. 1, the warehouse logistics management method of the present invention includes the steps of:
firstly, carrying out route design on an unmanned aerial vehicle according to a storage range and aerial photography parameters;
secondly, the unmanned aerial vehicle acquires aerial images according to the air route;
step three, carrying out image preprocessing on the collected image;
generating a three-dimensional model according to the preprocessed image;
generating DEM data and DOM data according to the three-dimensional model;
step six, forming a real-scene three-dimensional map of the storage area through data fusion of the three-dimensional model, the DEM data and the DOM data;
and seventhly, visually managing the warehouse logistics through the live-action three-dimensional map.
In this embodiment, the warehousing range in the first step includes a whole three-dimensional geographic space for goods storage and movement, and the aerial photography parameters include ground image resolution, image overlapping degree, altitude maintenance, photo inclination angle, and photo rotation angle.
In the concrete implementation, the ground image resolution is not lower than 0.08 m, the average course overlapping degree of the course overlapping degree is 70%, the average lateral overlapping degree is 55%, the altitude difference of adjacent photos on the same flight path is not more than 5 m, the difference between the maximum altitude and the minimum altitude is not more than 10 m, the difference between the actual altitude and the designed altitude is not more than 15 m, the inclination angle of the photos is not more than 8 degrees, the number of the photos exceeding 5 degrees is not more than 10% of the total number, the rotation angle of the photos is not more than 10 degrees, on the premise that the course and the lateral overlapping degree of the photos meet the standard requirements, the number of the photos not exceeding 20 degrees, and the number of the photos reaching or approaching the maximum rotation deviation angle limit difference on one flight path is not more than; the number of the photos with the maximum deflection angle in one photographic area does not exceed 10 percent of the total number of the photos in the photographic area.
In this embodiment, the unmanned aerial vehicle in step two acquires aerial images according to the route by using aerial triangulation, where the aerial triangulation includes encrypted point selection, encrypted point observation and calculation, and encryption process control.
In specific implementation, the photo data is used as original data of image control encryption, relative orientation is carried out piece by piece, absolute orientation is carried out according to an area network, integral adjustment of the area network is carried out by a light beam method, an encryption point result is obtained, a connection point of the encryption device is selected to be close to specified six standard points of 1, 3, 5, 2, 4 and 6, wherein the point of 1 and the point of 2 are selected to be within 1cm of an image main point, the point of 3, 4, 5 and 6 is consistent with a mapping point, and the distances from the azimuth line of the point of 3, 4, 5 and 6 are all larger than 4 cm.
In this embodiment, the encrypted point observation and calculation includes a relative directional tolerance and a model connection tolerance, the relative directional tolerance includes a standard point vertical parallax and a check point vertical parallax, the standard point vertical parallax is less than 0.003mm, and the check point vertical parallax is less than 0.007 mm; the limited difference of the model connection difference comprises poor plane position and poor elevation, and the poor plane position delta S is 0.5F
sm-3d, said elevation being poor
Wherein, Δ S is the poor plane position, Δ H is the poor elevation, F
smIs aerial photography scale denominator, d is photo baseline, f
lIs the focal length of the aerial camera.
In this embodiment, the image preprocessing on the acquired image in step three includes image enhancement on the low-illumination image and image stitching on the enhanced image, and as shown in fig. 2, the specific process of image enhancement includes:
step 3011, estimating an illumination component of the low-level image by using gaussian weighted bilateral filtering;
in the specific implementation, in the flat area of the image, the change of the pixel value of the image is small, the spatial weight plays a main role, namely, Gaussian blur is performed, while in the edge area of the image, the change range of the pixel value is large, and the similar weight is increased successively, so that the information of the edge is maintained. Therefore, the bilateral filtering can adaptively estimate the place with large edge difference of the low-illumination image, and effectively avoid the halo phenomenon while keeping the image edge information.
3012, performing adaptive nonlinear stretching on the saturation component of the low-light image in the HSV color space to obtain an enhanced saturation component;
in specific implementation, in order to make an image color more full, saturation needs to be stretched, R, G, B color channels exist in a color image, and the three channels need to be enhanced respectively when Retinex enhancement is performed on the color image, but three components in an RGB color space have strong correlation, and the correlation between the three components is damaged by the respective enhancement of the three components, so that image color distortion is caused, hue, saturation and brightness components can be separated in an HSV color space, and the original color structure of the image can be well maintained while the brightness and saturation of the image are improved.
3013, performing single-scale Retinex algorithm enhancement on the brightness component of the low-light image in the HSV color space to obtain a reflection component;
step 3014, taking an index of the reflection component to obtain an enhanced brightness component;
3015, combine the hue with the enhanced saturation component and the enhanced brightness component, and convert the HSV color space into an RGB space;
and 3016, performing contrast correction by using a global adaptive logarithm enhancement algorithm to obtain a final enhanced image.
In specific implementation, after the low-illumination image is subjected to Retinex enhancement based on bilateral filtering, the contrast improvement degrees of different images are different, so that the enhanced image needs to be corrected. The global self-adaptive logarithmic enhancement algorithm utilizes logarithmic characteristics and logarithmic mapping relations, logarithmic average brightness is introduced, when the dynamic range of a scene changes, no matter whether an image is too dark or too strong, the average brightness value is always smaller than or equal to the maximum brightness value, the display brightness value can be mapped to 0-1, smooth incremental increase of other display brightness values is guaranteed, the image cannot be distorted, the logarithmic characteristics are enhanced obviously for low-illumination images, and the logarithmic mapping relations are utilized to prevent the phenomenon of over-enhancement from influencing high-illumination images.
Under night or the condition of being shaded, there is luminance, contrast subalternation problem in the image that unmanned aerial vehicle shot, influences subsequent various image processing to a certain extent, consequently, improves image contrast, suppresses the dark space through image enhancement.
In this embodiment, as shown in fig. 3, the specific process of image stitching includes:
step 3021, processing the enhanced image by using an image matching algorithm;
step 3022, extracting homonymous feature points contained in an overlapping area in adjacent images in the same aerial photography zone and homonymous feature points contained in an overlapping area in adjacent images in the adjacent aerial photography zone in the sequence image;
step 3023, performing Euclidean reconstruction on the obtained homonymous feature points and performing triangularization in a three-dimensional space;
and step 3024, generating point cloud data of the warehoused goods through the sequence images of the warehoused goods.
In this embodiment, as shown in fig. 4, the specific process of generating the three-dimensional model according to the preprocessed image in step four includes:
step 401, point cloud calculation and registration
Solving transformation parameters among all frames of images, superposing and matching multi-frame images acquired at different time, angles and illumination intensities into a unified coordinate system by taking a common part of a storage scene as a reference for image registration, calculating corresponding translation vectors and rotation matrixes, and eliminating redundant information;
step 402, data fusion
Dividing the point cloud space into a plurality of small cubes through a volume grid;
in specific implementation, the registered depth information is still point cloud data scattered and disordered in the space, and only part of information of the scenery can be displayed, so that the point cloud data needs to be fused to obtain a more refined reconstruction model.
Step 403, surface Generation
And (3) generating a three-dimensional model surface by using a real warehousing scene picture acquired by the oblique image as a three-dimensional model map and texture mapping.
In this embodiment, the specific process of generating DEM data according to the three-dimensional model in the fifth step includes: and extracting characteristic points, lines and surface network construction interpolation according to the three-dimensional point cloud to generate the DEM.
When the method is specifically implemented, the DEM grid points are mapped on the three-dimensional model for inspection and observation, and each DEM grid point is ensured to be close to the ground; when the elevation error of the DEM grid point exceeds the limit and needs to be edited, a contour line and an elevation mark point with problems are searched for correction and measurement, or a characteristic point and a characteristic line are added, and the network interpolation DEM is reconstructed.
In this embodiment, as shown in fig. 5, the specific process of generating DOM data according to the three-dimensional model in step five includes:
step 501, introducing a space-three encryption result to establish a test area file and restore a model;
step 502, defining an operation area of a single model, generating an epipolar line image, and matching the epipolar line image to form a matching point and an equal parallax curve;
step 503, checking the matching result;
step 504, according to the elements of the inner and outer directions of the image and the resolution of the image, correcting and resampling the image by adopting a differential correction method to generate a single-model DOM;
step 505, adjusting hue and color;
step 506, setting an embedding range according to the outline coordinates, appointing a storage path, executing an image embedding command, and splicing into a whole DOM.
In specific implementation, after the image embedding is finished, the generated DOM needs to be checked, and the parts with blurred images and missing images in the edge area are repaired to prevent the image distortion of high-rise goods caused by the inconsistent projection direction.
The warehouse logistics management system of the invention comprises:
the unmanned aerial vehicle with the camera is used for aerial photography and acquiring a collected image;
a memory for storing a computer program for implementing a warehouse workflow management method;
a processor for executing the computer program in the memory;
and the display is used for displaying the real-scene three-dimensional map of the storage area obtained by the processor.
In specific implementation, the memory, the processor and the display can be integrated in various mobile terminals, so that the mobile terminal is convenient to carry and easy to manage.
According to the invention, the real-time three-dimensional map of the storage area is mapped by the unmanned aerial vehicle, so that the visual management of the warehouse logistics is realized, the change of the goods in the storage area is updated in real time, the specific position of each batch of goods is updated in real time after the goods enter the storage area, the target goods can be accurately and quickly found, meanwhile, the stock measurement and calculation can be carried out on various goods, and the intelligent management of the warehouse logistics is improved.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention, and all simple modifications, changes and equivalent structural changes made to the above embodiment according to the technical spirit of the present invention still fall within the protection scope of the technical solution of the present invention.