April 9, 2021
Defects that occur during manufacturing are not all equally relevant for the quality of the final product. For example, a small scratch on an automobile windshield is critical to the safety of the vehicle. In turn, a scratch of the same size on the plastic panel of the door is not significant.
If the manufacturer performs visual quality inspection manually, they run the risk of missing the important small defects. However, if a high-precision automated solution is applied for surface inspection, there is a problem that even the parts with irrelevant small defects are classified as defective and taken out of production. Some manufacturers estimate that if all parts that have a fault are considered defective, the company will have to dispose of around 50% of the parts. Yet, for these producers, a defect of size starting from one centimeter is relevant to the quality of the product.
FotoNow, in addition to the main objectives of improving the quality and productivity in the manufacturing process, also focuses on sustainability. Our solution is designed to allow the manufacturer to optimize the quality assurance process while reducing waste.
To maximize the value of FotoNow and deliver customer-centric performance, we have developed a “Human-in-the-loop” mechanism. After the Deep Learning-based model is trained, our solution performs automatic quality checks in the complete field of vision of the camera. To ensure maximum quality assurance, FotoNow detects all existing defects on the image. In the surface inspection process, errors starting from micron-sized are identified. In the next step, the employee is notified of the error. Accordingly, the employee has the opportunity to look at each case individually and assess whether the defect is indeed significant. Non-significant flaws can be noted as such with one click in the FotoNow application. The model then learns on its own what size, depth, shape, or category of defect can be classified as irrelevant for specific parts. After that, the algorithm makes decisions based on newly learned knowledge.
No IT knowledge is needed for model retraining. Any employee can make changes following the suggestions in the application with intuitive user interface.
This way, the manufacturer can be sure that all the substantial faults are detected and, at the same time, the products with non-significant defects are not disposed in vain.
All defects are detected and shown on the screen.
Choose areas you want to mark as faulty or faultless.
The algorithm learns automatically and makes further inspections considering your corrections.
FotoNow is a quick learner and can understand the requirements after just a few retraining sessions. With our solution, the manufacturer gains a smart tool that supports the employees in quality control and optimizes quality management.
Please, contact our team to request a demo or learn more about our self-learning visual quality inspection technology.
One of the latest technology trends for enterprises is the combination of Artificial Intelligence and the Internet of Things. Merged, the new trend...
As with all tools based on Deep Learning, the performance of solutions for automated visual quality inspection depends directly on the quality of...
AI technologies are changing our way of living and working. Enterprises are actively looking for innovation through AI. The benefits of...