Menu
Kitov’s technology focuses on the automated planning of quality inspection, the quality inspection itself, and the synergy between the two. To perform high-level quality inspection, especially for complex parts and products, the component must be examined and inspected from several precise 3D poses . Aside from the ability to accurately control mechanical systems such as robotic arms and cameras, it is imperative to determine the optimal 3D poses for each specific inspection task.
Using the Kitov platform, QC engineers can define their inspection requirements directly on a complex CAD model – simply and intuitively – saving days, weeks, or even months compared to manual programming and finetuning.
Particularly useful for parts with complex geometries and parts with high-mix and customizations, Kitov’s software automatically extracts the specific geometric, PMI, and other semantic information defined on the CAD for each inspection requirement. This information is passed on to the relevant semantic detectors performing the inspection tasks, such as the surface detector, label detector, screw detector, etc.
The ability of Kitov’s platform to provide the inspection detectors with a wealth of information about the inspected component, including its precise geometry, available PMI, and other semantic information that exists in the CAD, allows the detector to maximize its inspection performance. This applies to both the Kitov pre-installed detectors and any third party using Kitov’s open platform to develop specific and customized detectors.
Automatic Inspection Planner
Kitov’s planner integrates with existing or new automation and robotic systems to generate an optimized test plan. The test requirements are defined on the CAD system using Kitov’s award-winning CAD2SCAN technology and best practices for each test requirement, including the best illumination angles and imaging parameters. The software optimizes the inspection plan for accuracy and throughput, guaranteeing full coverage while reducing robot motion and the number of images needed to cover all regions of interest, thereby minimizing the total inspection time.
Kitov’s planner optimizes the inspection plans not only for Kitov systems, such as the Kitov Core+ – but also for customized configurations that consist of different hardware and robots, including different robot vendors, robot sizes, number of joints (DoF), etc. This is done by specifying the system hardware, the CAD of the inspected part/product, and the required quality tests to perform. The outcome of the planner is an optimal and safe inspection plan that will control the robot’s path and the acquisition of high-quality data. The plan is safe in the sense that it will guarantee no collisions with the part or objects within the environment.
The planner will optimize for an inspection plan that satisfies best practices and imaging instructions for each test type, such as the number of images to acquire, the optimal 3D poses of the camera (or other tools), and the illumination relative to the component to be inspected. The best practices can also specify parameter tolerances, whether to group tests, dictate the preferred illumination, and request that the planner perform an appropriate illumination calibration, among other behaviors.
The planner supports the online alignment of parts that might be handed to the system with geometrical errors. In such cases, the actual test plan will be altered, during the inline inspection, to align with the accurate position of the part. The system’s camera determines the actual 3D position of the part without the need for external trackers.
A useful example of the automatic inspection planner is when inspecting non-planar surfaces, for example, a blade of a turbine engine. In this case, every small surface patch has its optimal viewing and illumination angles that allow for the best image where defects can be detected. The planner’s challenge is to ensure maximum image coverage with a minimal number of images so if a surface defect exists, it will be visible in one of the images. The planner must also with material properties and reflections caused by illumination.
The synergy between inspection plan and the inspection itself is Kitov’s strongest advantage. We specify the best practices for each type of inspection, determine the ideal imaging poses and parameters, and select the right algorithms to use accordingly. Another advantage of Kitov’s solution is our open software platform. We concentrate on technology and our partners focus on system building and customization for specific markets and products.
Kitov’s main technologies follow a specific quality inspection process:
The CAD2SCAN defines and marks the inspection requirements directly on the CAD system easily and intuitively.
Kitov’s automated planner optimizes the quality test plan to run in the most efficient manner, exploiting any robotic and automation system while considering the best practices for each test type.
The AI-based visual quality inspection operates in synergy with the planner and uses a hybrid approach that combines classical Computer Vision and Image Processing methods with AI and Deep Learning networks for best inspection results that are constantly being enhanced.
All Kitov’s technologies are packaged in an Open Software Platform, which enables Kitov’s customers, partners, and third parties to extend and customize the quality inspection capabilities to their specific needs, including specific robots, tools, and software detectors.
Kitov’s open software platform is designed to easily integrate with any automation environment and easily define and perform an inline quality inspection.
This is achieved with several interfaces that allow connectivity with:
Fully automated visual inspection solutions for a broad range of production lines and markets.