In the AEC industry, Machine Learning has become a prominent factor that is making things easier and smooth. Using Machine Learning, Scan to BIM workflow can easily achieve higher accuracy with nominal human involvement. By leveraging the supremacy of machine learning, a modeler can get an automatically generated BIM model. Scan to BIM-based on Machine Learning includes laser scanning, point cloud segmentation and ordering, and final conversion into BIM design in the workflow. But in this workflow, for a particular output, it is required to follow a few key points and mandatory to avoid a few mistakes. Generally, mistakes that happened during scanning, conversion, and point cloud segmentation are like not performing proper scanning, not checking all the defined key points while conversion, and can be improper segmentation or ordering. All these mistakes can affect the final modeling output. You can consider the following points as best practices to follow while processing Automated Scan to BIM workflow.
Let’s see the step-by-step workflow with beneficial assistance.
It is extremely significant to follow standards that are set for exact scanning. Scanning is done on-site and there are a lot of chances to happen mix-ups and mistakes. It is mandatory to convey all necessary details to the field staff so that scanned data can be easily converted into a BIM model. So, let’s see the things to avoid during Scanning:
Visiting a field without a checklist
It can create a lot of confusion and of course, can waste time. Hence, create a step-by-step checklist of required factors before visiting the field.
Not defining the exact level
The level depends on the choice or requirement which you like to perceive in Scan. Defining the level of Scan can make BIM conversion clearer.
Use of normal Laser Scanner
The highest quality laser scanners are the only good choice for Scanning. Details scan takes minimum time to convert into a model. Hence, always use the highest quality scanner which can offer the smallest details.
While scanning and visiting the field, there are a few set standards and procedures which need to be followed. Make proper documentation with the help of a construction designer.
Point Cloud Segmentation and Ordering
Point Cloud Segmentation is the main step in Scan to BIM which includes the separation of point cloud data as per the entities. Better point-to-cloud data can make a perfectly aligned BIM model design. But there are a few standard practices that every modeler has to follow for accurate results and more importantly the mistakes that one should avoid.
So, here are the mistakes to avoid in Point cloud segmentation and ordering:
Segmentation of the point cloud data is important for designing separate entities. It is based on the entities in point cloud data. This can be also done by comparing data with the original photos and videos of the site. Hence, categorize or separate point cloud data properly to avoid inappropriate ordering and designing in Scan to BIM.
After segmentation of point cloud data, it is needed to order the entities in specific alignment or flow. Ordering can be in numbers, types, or related to site structure. So don’t miss the ordering part to avoid defects in the BIM model. By creating a sequential order of the entities, the modeler can easily convert Scanned data into a BIM model.
Conversion is the final phase of the Scan to BIM. It is perhaps one of the less critical steps in Scan to BIM workflow. But there are so many obstacles and difficulties in the process. It is always good to follow standard procedures and to avoid mistakes in the process. This helps to save time and make the design more remarkable. Here are the mistakes to avoid and things to follow while converting the Scan into BIM.
Rework/ converting Scan without checking details
It is a must thing to ensure that received scans are stick to a defined level of detail which are specified in the procedure. Don’t start the converting process only by transferring the scan data into a specific drive. It can waste your time in rework.
Improper upload of data
Upload scanned data onto any suitable 3D design software in a proper way. Save or upload data in an organized fashion by following the numerical order of the scans.
Inappropriate segregation of levels
Scans have millions of data points in the point cloud which is uploaded in the design software. It can create confusion if it is not separated into the proper levels. It is convenient to make diverse stages of elevation and set view depth as per the original plans.
This is enough to make a perfect BIM design. By following all these suggestions and by avoiding these mistakes, anyone can get a proper result out of the Scan to BIM workflow.
Things to do while executing ML-based Scan to BIM workflow
Decide the purpose of Data
While planning the scanning, it is a must to decide the purpose of the Scan. Modeling can be the only purpose but by confirming the criteria it can be more operative.
Enter the parameters
While sending someone on the project site, it is required to convey the details like a number of measurement points and scan density. Don’t forget to enter these parameters because they are essential for faultless scanning.
Take multiple scans
To get a complete illustration of the site or project, it is helpful to take several scans from different locations. Hence don’t fail to take multiple scan copies of the site.
Transfer data from the scanner to any drive or platform. But try to use better software for registration and analysis before modeling.
After saving or transferring data into the cloud, it can be imported into any modeling software. Don’t lower the quality and precision while modeling.
Workflow of Scan to BIM using Machine Learning demands exact inputs and processes to avoid defects in outputs. We hope that now you have better awareness about the mistakes to avoid in the workflow. Hoping that all the suggestions will help you to maintain the workflow of Scan to BIM. Else feel free to connect with us for better assistance and suggestions.