dragonfly-science / csm-test

0 stars 0 forks source link

please note I am replacing "pipeline" throughout #8

Open katrin-berkenbusch opened 3 years ago

katrin-berkenbusch commented 3 years ago

To avoid jargon.

xycarto commented 3 years ago

@katrin-berkenbusch

'Pipeline' can refer to two different items.

  1. The jargon developers like to use to describe work flows
  2. A specific operation used by PDAL. PDAL uses a command pipeline to refer to a pipeline JSON document. This is the PDAL syntax. The JSON document contains all the steps necessary to complete an operation
katrin-berkenbusch commented 3 years ago

Thanks, so maybe this aspect needs to be distinguished in the text? Is there a good non-jargon term to refer to the PDAL pipeline?

xycarto commented 3 years ago

JSON config? Although tthe pipeline doc is not really a config either. I've just always known it as the 'PDAL pipeline'. Let me have a think about this

katrin-berkenbusch commented 3 years ago

thanks

On 30/06/2021, at 12:38 PM, XYCarto @.***> wrote:

JSON config? Although tthe pipeline doc is not really a config either. I've just always known it as the 'PDAL pipeline'. Let me have a think about this

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

katrin-berkenbusch commented 3 years ago

Still no solution for "pipeline"??? @xycarto

xycarto commented 3 years ago

@katrin-berkenbusch

I had a good read of the PDAL documentation to look for a good alternative for pipeline. Pipelines in PDAL are a JSON representation of a workflow PDAL uses to complete an operation. The difficulty in finding an alternative term is that PDAL, throughout their documentation, defines to these JSON documents as Pipelines.

In the report we are using the term 'pipeline' as PDAL documentation would use it; however, it could be helpful to replace term as something more descriptive. Perhaps we can define pipeline early like:

"PDAL implements its processing workflows through JSON documents called 'Pipelines'. Pipelines represent the operations PDAL will following in reading, writing, and translation of point cloud information."

We are using 'Pipeline' documents for our operations; however, PDAL can operate without these JSON documents and operate through a series of commands.PDAL operations. Pipeline may be necessary in a few locations to stay consistent with PDAL documentation, but alternatives to terminology like using, commands, operations, methods, or processing steps, when describing what is actually happening can be more helpful.

I've gone through the report and tried to develop some examples where this can be replaced:

B.3.2 "This method for developing specific configurations for processing individual point clouds gives us the control needed to run these processes across myriad cores"

4.1 (paragraph one) "The CHM.fly, generated from unclassified LiDAR point clouds using a PDAL configured processing method, corresponded closely with the other CHMs analysed for the test region."

4.1 (third paragraph) "A significant advantage of the process developed here is its scalability: incorporating PDAL allows for the development of CHMs from unclassified point clouds, removing vendor quality limitations (ground point filtering), and PDAL processing configurations can be readily incorporated into the AWS infrastructure."

3.1 (paragraph two) 'The point cloud filtering and labelling processes for the DTM generation from OpenTopography and Dragonfly revealed that the PDAL classification operation classified 25.1% more ground points than the CHM.topo, increasing the resolution from 3.70 points m −2 to 4.63 points m −2 ."

I cannot really comment on the instances outside of these where @bogdanstate uses the 'pipeline' terminology