Visual inspection in manufacturing using artificial intelligence

Two Uses of Artificial Intelligence Visual Inspection: Fruit and Ceramic Tile Quality Control Defect Detection

Visual inspection by computer vision is a subclass of the set of techniques and phenomena introduced under the umbrella of artificial intelligence. As its name suggests, it is about the development of the ability to see in machines , which can be considered close to the human perception of analyzing objects through vision. Its applications include object detection, classification, transformation, image and video processing, all to solve real-world problems. One of our recent challenges is based on the application of this technology.

Below we describe two applications of visual quality inspection in processes using artificial vision techniques. This type of application aims to reduce the number of operators who carry out visual inspections to identify defects in manufactured products, making their work easier and faster.

  1. Visual inspection of the quality of ceramic tiles

Construction material companies, including ceramic tile manufacturers, need to carry out visual inspections of their products.

By implementing machine vision algorithms, these manufacturers can save a lot of time and money. The accuracy and speed provided by AI-based visual inspection cannot be matched by any other technique.

These inspections identify defects such as bulges, depressions, color casts, or unwanted lines. In addition, vision system images allow automation engineers to further fine-tune the detection of acceptable or unacceptable aesthetic defects.

It takes a lot of time to formulate rules to consider all the possible visual defects that could occur during an automatic inspection:

It is necessary to identify any type of defect, to determine its patterns, colors, textures and eventually its values, grades, usability and sale prices.

Not every imperfection caused by flaws or color irregularities is grounds for rejection. Some are within the acceptable range.

It is best if defects are detected before the clay dries, the material can be recycled, and waste is eliminated.

Just as a car with autonomous driving incorporates different sensors (cameras, radar, ultrasonic sensors) to interpret reality through the treatment of the images provided, combining this information, it is possible to use different sensors and technologies to automate the quality control of Ceramic tiles:

  1. Image cameras and OpenVINO

To solve these inspection challenges, visual inspection algorithms combined with an in-line scanning camera and LED lighting can be implemented.

The hardware must be up to the task:

  • High-performance CPU to enable tile inspection at speeds of up to 30,000 pieces/hour.
  • Extensive I/O interface with support for multiple cameras at an image resolution of at least 3840 x 2160
  • Fast GbE interface for fast networking.

Software-wise, there are various tools and algorithms available, such as:

  • Intel’s OpenVINOOpen Visual Inference and Neural Network Optimization embedded deep learning toolkit along with Myriad X VPUs for vision processing, which offers seamless software/hardware integration with the Intel Movidius engine, and includes model optimizer, the inference engine, pre-trained models, machine vision libraries, code samples, and other tools, or

Cognex ‘s VisionPro Deep Learning System , a proprietary deep learning solution, identifies multiple cosmetic defects on the original plate based on a small sample of around 100 approved images. Inspecting a 200 x 32 cm material takes about two seconds.

  1. X-Ray and Detectron2

Another possibility is to use x-rays. Limitations in resolution and sharpness of boundaries and interfaces in images reconstructed from data collected by X-ray computed tomography (CT) make it difficult to extract and segment features of interest.

To avoid this, the Faster RCNN module of the open source object detection algorithm Detectron2 developed by Facebook’s artificial intelligence team is used. This approach employs multiple convolutional neural networks configured in series for real-time object detection to identify defect positions (such as cracks) in two-dimensional slices of the three-dimensional tomography dataset, capable of producing detection results with superior accuracy. at 85%.

To train the detection algorithm, the user must manually incorporate several images to use in the training processes.

  1. Low cost spectral radiation

Another newer possibility is the implantation of terahertz cameras, which emit a non-ionizing, innocuous radiation, and allow the ceramic pieces to pass through and be able to measure their apparent density, allowing the inspection of the tiles, so that the same density can be maintained in all parts, achieving high production quality and significant cost reduction.

In addition, the application of these advanced optical sensors has shown that they can also be used to detect the presence and concentration of the inks applied in the decoration of the tiles, without having to wait for the material to be baked. Therefore, they can be used to detect flaws in part decoration and optimize the printing process.

Related Posts