Announcing OpticVyu Artificial Intelligence Based Image Processing Solution

Construction site automation is the need of the hour. When site imagery tunneled through artificial intelligence modules yields outstanding results. And the data thus retrieved is most valuable for any construction management & planning team.

OpticVyu Artificial Intelligence Module processes images captured by OpticVyu construction cameras & detects workers’ PPE to check the safety measures followed at the site thus minimizing the manual efforts to ensure employee’s safety.

Image processing is a cost effective version of video processing which minimizes the technology cost & makes it affordable for every construction project. This also eliminates the need of the dedicated LAN based internet connection to upload huge data to cloud servers. Having said that, the output derived from HD images’ processing is also precise as compared to processing an h264 video stream.

What is the PPE Detection Module?

OpticVyu AI module detects PPE & checks if a person is wearing a hard hat or not. The detected people will be categorized into three groups viz. 

  • person with PPE,
  • person without PPE,
  • person detected but unrecognized for PPE.

OpticVyu Proprietary PTZ Image Processing Camera Solution

Image processing invariably depends on the size of objects captured in an image & its resolution. With any fixed position camera, images can be captured at a predefined orientation & persons captured are either close or far away. Sometimes, the pixel size of a person seems very small, so that image processing becomes almost impossible.

Problem statements:

What if the people are not captured in the defined image frame?

What if the detected people are not clear enough to be processed by the AI module?

Solution

OpticVyu proprietary PTZ camera-based multi-positioned varying zoom capturing solution ensures the image capturing in all possible directions with up to 45x zoom-in capabilities to yield the best results for image processing. Each site is dividing into 25-50 presets (predefined viewing locations) to collect as many images as possible with utmost clarity. The camera rotates, zoom-in to different preset locations & captures images every 1hr & processes for PPE detection.

Requirement to Setup a PTZ-AI Camera

We prefer to install the camera at the center of the site for maximum coverage with just one camera. The camera operates on an inbuilt SIM-based cellular network. External Wifi is not needed. A regular household power supply (230V-50Hz) is sufficient to run the camera.

Dashboard Features for Image Processing

The image processing dashboard is divided into two categories viz. a) Graphs, b) AI-data

Graphs give instant information about the people detected with or without PPE in any selected time duration. Whereas, AI-data option gives a detailed information of all the people detected with or without PPE on a given date for the selected present location. Refer to the below videos:

PPE Reports

Getting daily & weekly PPE detection reports is super easy with OpticVyu. Configured email ids receive daily or weekly reports with the summary of detected people. With automated reports, follow the site safety measures through email updates even when you’re busy with other important works.

Can Existing OpticVyu Construction Cameras be Upgraded for Image Processing?

Yes! Existing OpticVyu cameras can also be upgraded for image processing within 5mins. As soon as the integration is activated, all the new images will be processed for PPE detection. As mentioned earlier, the existing cameras might be capturing a large site area & the pixel size of the detected person can be too small to process for PPE. Thus not all existing cameras can yield the desired output through this integration.

Does Time-lapse & Other Construction Camera Feature Work as It Is?

Yes! May it be an OpticVyu PTZ-AI camera solution or the integration with the existing camera for image processing, all the features like time-lapse, reports, comparison, integration with BIM, Procore, project planning work as it is.

Limitation & working range

Although the OpticVyu image processing module delivers one of the best output in the construction camera segment yet some practical limitations restrict the solution to report 100% accuracy:

  1. The pixel size of detected people should be more than 5% of the shorter image dimension. For example, if the image pixel size is 2560×1440, the smallest detectable person size should be at least 1440×0.05 = 72 pixels.
  2. For the same zoom level & frame orientation, a higher resolution camera produces better results.
  3. A fixed position camera produces the best results for the people captured within 15 meters.
Default image
OpticVyu
OpticVyu Operations is the leading construction time-lapse, 360-degree image-based interior monitoring, material tracking, service provider.
Articles: 25