LIDAR for Bulk Sugar volume Measurement

Hello,

One of my clients needs to measure bulk sugar volume on a daily basis in a large (70m x 70m x 15m) indoor warehouse.

What comes to mind for this problems in a Lidar based scan of the bulk surface, from which the volume can easily be recalculated. It may or may not be useful/desirable for the Lidar to be on a drone (closed space makes it tricky, and not too large area so fixed installation could be ok).
Has anyone seen this kind of problems solved by Lidar equipped solutions.
Or has any other suggestions of directions to explore?

Thanks a lot.

sugar.jpg

You need to be a member of diydrones to add comments!

Join diydrones

Email me when people reply –

Replies

  • Hi Govert, thank you for your reply, will PM you.

    Best. Vincent.

    Govert van Drimmelen said:

    Hi Vincent,

    Our company - Ronin - makes industrial laser scanners for exactly this kind of real-time bulk inventory application.

    3702382886?profile=original

    Lots more info, brochures etc. on the website at http://www.roningms.com.

    You can also PM me and I can get a sales manager to contact you.

    Regards,

    Govert

  • Hi Vincent,

    Our company - Ronin - makes industrial laser scanners for exactly this kind of real-time bulk inventory application.

    3702382886?profile=original

    Lots more info, brochures etc. on the website at http://www.roningms.com.

    You can also PM me and I can get a sales manager to contact you.

    Regards,

    Govert

  • Thanks Rick.

    Rick F said:

    I worked with a minerals company that used a reflectorless theodolite  set up at 4 known stations around an indoor mineral stockpile with a volume of around 15,000 cubic metres.  Multiple readings from top to bottom along several different lateral angles were taken from each station.  Processing of the resulting data file was done with standard surveying software (I forget which one).  Results were accurate enough for our purposes.  This was done several times a month as a check on production and outgoing shipments.

  • I worked with a minerals company that used a reflectorless theodolite  set up at 4 known stations around an indoor mineral stockpile with a volume of around 15,000 cubic metres.  Multiple readings from top to bottom along several different lateral angles were taken from each station.  Processing of the resulting data file was done with standard surveying software (I forget which one).  Results were accurate enough for our purposes.  This was done several times a month as a check on production and outgoing shipments.

  • Or just install multiple cameras set to trigger at the same time and process in pix4d. Sort of like their tower based software version.

  • Hi,

    This looks exactly like what we are trying to accomplish.

    1. Totally agree, several sensors needed to cover shadow areas (although warehouse not too low, so it should make the coverage easier.

    2. There won't be any machinery in the warehouse so it will make automation easier

    3. Why do the lidars need to be gimbal mounted, don't they cover a sufficient spatial volume (placed upside down on the ceiling for example).

    Your former services look very relevant, can we PM to continue the conversation.

    Best regards,

    Vincent.

    Laser Developer said:

    Hi Vincent,

    We used LiDAR for volumetric and mass estimation of stock piles for many years, including on sugar stockpiles. These were permanent, static installations where the customer required hourly or daily updates. Whilst we're not directly in this business anymore (we sold it to an F500 company), our standard laser modules are capable of generating the required point clouds if you want to put them onto a gimbal or servo mounting.

    3702379591?profile=original

    There are a few tricks to keep in mind:

    Firstly, no matter which technology you are using, there is always the possibility of getting "shadows" where the material cannot be seen from the perspective of the sensor. This is especially true inside a low warehouse where you can't get a vertical, birds eye view from far enough away.

    The solution is to use multiple devices looking from different directions. The image below was made using 6 LiDARs mounted high up against the walls of the warehouse (60m x 20m). There was no room in the center due to an overhead conveyor belt. 

    3702379531?profile=original

    The second issue is that fully automatic rendering of the 3D image may not be practical. This is because there is often machinery, conveyor belts, trucks and other large objects that move around the stockpile, changing the surface profile. If these are not removed correctly then you can get inaccurate results. In the stockpile below there are two feed conveyors so we used two LiDARs viewing the pile from opposite sides and mounted on the local superstructure. The conveyors were removed manually before calculating the results.

    3702379610?profile=original

    In closed environments where there are clearly defined boundaries to the stockpile, such as in a silo, the situation is much simpler and very accurate results can be obtained. Below is an image from a sugar silo:

    3702379669?profile=original

    And here is a grain silo with some measuring markers added for visual effect (the scale is in tons):

    3702379545?profile=original

    The biggest advantage of using LiDAR is that lighting conditions don't matter and, as you have rightly said, getting enough features to do accurate photogrammetry is quite difficult in a typical industrial, warehouse setting and impossible inside closed silos without additional lighting. We ran this LiDAR scanning business as a service where the customer paid for each scan (image and data) and either bought or leased the equipment. Everything was done online so that our rendering software could be continually improved without the customer having to buy it and the LiDAR hardware was relatively simple and inexpensive.

  • Hi Chris,

    Yes indeed lack of features on the raw sugar stockpile.

    Thank you for the pointer on the Leica BLK360, amazing piece of portable and well packaged technology, and I can imagine quality post-processing with Autodesk Recap.

    The sugar stockpile will probably require a more industrial setup though, but thanks a lot for the introducing that interesting product.

    Best.

    Chris Anderson said:

    Fair point about the lack of visual features in sugar. You might want to check out the new low-cost Leica BLK360 laser scanner/camera, which is designed to work with the same ReCap engine that we use for photogrammetry and should give quite accurate results. 

  • Hi Vincent,

    We used LiDAR for volumetric and mass estimation of stock piles for many years, including on sugar stockpiles. These were permanent, static installations where the customer required hourly or daily updates. Whilst we're not directly in this business anymore (we sold it to an F500 company), our standard laser modules are capable of generating the required point clouds if you want to put them onto a gimbal or servo mounting.

    3702379591?profile=original

    There are a few tricks to keep in mind:

    Firstly, no matter which technology you are using, there is always the possibility of getting "shadows" where the material cannot be seen from the perspective of the sensor. This is especially true inside a low warehouse where you can't get a vertical, birds eye view from far enough away.

    The solution is to use multiple devices looking from different directions. The image below was made using 6 LiDARs mounted high up against the walls of the warehouse (60m x 20m). There was no room in the center due to an overhead conveyor belt. 

    3702379531?profile=original

    The second issue is that fully automatic rendering of the 3D image may not be practical. This is because there is often machinery, conveyor belts, trucks and other large objects that move around the stockpile, changing the surface profile. If these are not removed correctly then you can get inaccurate results. In the stockpile below there are two feed conveyors so we used two LiDARs viewing the pile from opposite sides and mounted on the local superstructure. The conveyors were removed manually before calculating the results.

    3702379610?profile=original

    In closed environments where there are clearly defined boundaries to the stockpile, such as in a silo, the situation is much simpler and very accurate results can be obtained. Below is an image from a sugar silo:

    3702379669?profile=original

    And here is a grain silo with some measuring markers added for visual effect (the scale is in tons):

    3702379545?profile=original

    The biggest advantage of using LiDAR is that lighting conditions don't matter and, as you have rightly said, getting enough features to do accurate photogrammetry is quite difficult in a typical industrial, warehouse setting and impossible inside closed silos without additional lighting. We ran this LiDAR scanning business as a service where the customer paid for each scan (image and data) and either bought or leased the equipment. Everything was done online so that our rendering software could be continually improved without the customer having to buy it and the LiDAR hardware was relatively simple and inexpensive.

  • 3D Robotics

    Fair point about the lack of visual features in sugar. You might want to check out the new low-cost Leica BLK360 laser scanner/camera, which is designed to work with the same ReCap engine that we use for photogrammetry and should give quite accurate results. 

  • Hi Chris,

    Thank you for your fast reply.

    Yes photogrammetry can be an option, and I am familiar with open mine or quarry volume measurement using drone-based data and photogrammetry.

    Nervertheless, I have the feeling that a heap of sugar in and indoor warehouse with possibly not a good lighting would not be as good a candidate for photogrammetry, which relies on reverse calculation from different perspectives of elevations and volumes: in the sugar bulk, no defined edges, not much difference of colors, uniform lighting etc..

    Whereas a Lidar, whatever the shape and lighting of the object, measures distances in its range and target directions, so gives what we need  for volume calculation. Not as good as photogrammetry for textures, but its primary data output is distances in a direction therefore point cloud therefore volume.

    In any case, data processing can always be given to sitescan/pix4d if they are able to handle the data format.

    Cheers.

    Chris Anderson said:

    It's actually a pretty easy job to do with photogrammetry. Put cameras in the right places and send the data to ReCap or Pix4D. Here's how we do it with drone data in Site Scan:

This reply was deleted.

Activity