Machine Vision Inspection
PAQi allows for the use of traditional machine vision with an API to National Instruments’ Vision Buider Automated Inspection and AI-based inspection or a combination of both to provide a manufacturing tool faster and more accurate than human inspection. While traditional machine vision excels at physical measurements based on a rules-based configuration, AI-based inspection utilizing machine learning to determine the rules required to do reliable inspection from example. Furthermore, to speed the integration of visual inspection, PAQi provides a means for getting started quickly with traditional machine and transitioning over to AI-based inspection when enough samples have been collected to improve the inspection accuracy.
PAQi's AI-based vision inspection system employs cutting edge deep learning pipelines, enabling it to perform a diverse array of inspection tasks with unparalleled accuracy and efficiency. PAQi excels in object/defect detection, classification, and segmentation across various manufacturing industries and applications.
-
Provides methods for deploying localized inference on trained models produced by multiple online platforms including but not limited to Roboflow, Vertex AI, and QC Hero.
-
Advanced users are able to train and deploy their own custom proprietary models in parallel to standard open-source trained models.
-
By localizing inference on customer provided hardware or our PAQi controller, uptime is maximized by avoiding pitfalls associated with cloud-based inference such as phoning home. For applications requiring a high detector count, up to 30 object detectors can be run simultaneously using the QC Hero training platform.
-
Our patent pending calibration methods provide a quick no-coding means for turning predications into action whether it be signaling a robot to place a defect part in a reject bin, stopping an extruder from producing waste, or actuating a paddle to divert bad parts on a conveyor line.
Structure
The diagram below explains the overall architecture of the PAQi system. All connections represent Ethernet traffic. Because all connections are Ethernet based, the individual system components can reside on a single or multiple Windows based computers.
Features
Classification
PAQi utilizes cutting edge deep learning to accurately classify objects based on predefined criteria. By analyzing image features and patterns, PAQi efficiently categorizes items with precision.
Object Detection
PAQi employs sophisticated object detection algorithms, leveraging cutting-edge computer vision techniques to identify and locate specific items within images. Through robust feature extraction and analysis, PAQi efficiently detects objects with precision, facilitating seamless integration into manufacturing workflows for comprehensive quality assurance.
Semantic Segmentation
Similar to object detection with a different model architecture, segmentation is able to assign classes to objects within and image to provide a means for differentiating object for defect detection and part presence verification.
Measurement
PAQi harnesses traditional machine vision methods for precise measurement tasks, complementing its advanced object detection capabilities. By integrating measurement functionalities with object detection, PAQi delivers comprehensive inspection solutions that ensure accurate sizing and dimensional conformity in manufacturing processes
Multi-Region Inspection
PAQi facilitates inspection of multiple regions of interest within an image simultaneously, increasing efficiency and enabling comprehensive analysis of complex or large-scale components.
Color Inspection
PAQi enables color inspection and analysis, allowing for the detection of color variations, inconsistencies, or defects in products or materials, ensuring color consistency and quality standards.
Pattern Recognition
PAQi incorporates pattern recognition capabilities to identify specific patterns or textures within images, facilitating the detection of defects or anomalies that may be indicative of manufacturing issues.
Parts Counting
PAQi offers parts counting functionality, allowing for accurate counting and enumeration of individual components within images. This feature is particularly useful for inventory management, production monitoring, and quality control processes, ensuring precise tracking and management of parts throughout manufacturing operations.
Defect Detection
Object detectors, classifiers, and/or segmentation can be used to find defects in manufactured parts including but not limited to dents, scratches, voids, discoloration, debris, or surface flaws.