KlinterAI
Klinter, the name, originates from Clinker, a residue formed from Manufacturing Construction Concrete.
Class Balance
Dataset Description
From about 700+ images, 77 classes are identified with at least 1 object from each class. The Dataset is custom built for Construction Foundations Site. The dataset is collected from the year 2000 onwards from Construction Sites using Digital Cameras. The tasks undertaken by the dataset requires us to collect more images related to such Construction Sites using Videos and Internet Images. The Dataset is available at:
KlinterAI Roboflow Universe Dataset
Hydrofriase Cutter
Activity Recognition
The outcome of a recognized activity is a block of text generated from an image and the objects detected. Such recognized activities compose together to form the BIM (Building Information Modeling) activities in a BIM 4D model. The BIM 4D Model is connected to the BIM 5D Model using the Conceptual Classification Codes presented by Revit which each activity exposes.
LOAD CARRYING CRANE WITH MOVING TRUCK
Object Detection Prediction
Computer Vision is found to be useful for producing the Scheduling of tasks from a Construction Site. Using Vision as a sensor from Depth Cameras, the detected objects form an input towards the recognition of activity. The Core application will run an Object Detection Model and an Activity recognition Model. Advanced Computer Vision techniques such as OpenVINO are used for Performance Improvement.