Image-based quality control “Damage Detection”

Automated inspection of damage and defects using AI-assisted quality control of reflective surfaces.

Where does the AI application offer the greatest benefit?

The inspection of reflective surfaces, for example of quality defects in production or damage to vehicles, is often still carried out manually and therefore turns out to be time-consuming and requires a high level of expertise on the part of the persons in charge. In order to increase the productivity of the processes, “Damage Detection”, an AI system for the quality control of shiny or diffusely reflecting surfaces, has been developed. It is suitable for a wide range of applications: from industrial production, for example in the automotive industry, to the leasing and insurance industry and to automotive appraisals. The system works automatically and takes less than a minute for surface inspection. The quality defects found are categorized using Deep Learning. The combination of deflectometry, i.e. the contact-free detection of reflective surfaces, conventional image recognition processes and artificial intelligence methods is a unique system.

What are the quality indicators of such AI applications?

  • “Damage Detection” offers low hardware and maintenance costs as a solution. The AI application can be used flexibly and retrofitted into existing productions. Among other things, the solution is able to work under the influence of stray light (e.g. ceiling lighting in a production hall).
  • The system delivers 100% test coverage. It combines high accuracy with low hardware requirements. Currently, defects as small as 0.1 mm can be detected on 1 m component size.
  • By using Artificial Intelligence, the system can be trained to detect various types of defects on the surface.
  • In addition to highly reflective surfaces, the latest developments also allow diffusely reflective surfaces to be tested for their quality.
“The combination of conventional image recognition
methods, AI methods and the detection of reflective
surfaces is unique.”
Dr. rer. nat. Theresa Bick
Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS

What AI technology powers the demonstrator of KI.NRW?

Computer Vision

The solution relies on Convolutional Neural Networks (CNNs) for detection and classification of the surface anomalies. The CNN model belongs to the Artificial Neural Networks, which are suitable for image recognition and understanding.

Informed Machine Learning

By incorporating expert knowledge, the system is able to make dataintensive neural networks practical for industrial production. The amount of training data and the annotation effort can be kept comparatively low.

Combination with classical image processing

The AI is linked to classical image processing. The system uses the advantages of both worlds – fast, approximate algorithms from classical image processing and the powerful methods of deep learning.

What does the AI demonstrator show?

Quality control of shiny or diffusely reflecting surfaces is notable for its simple, mobile hardware design. It operates independently of ambient light and is fully automated. It takes less than one minute for a surface inspection.

Where is more information to be found?

Image recognition in practice

Many “good practice examples” as inspiration for the possible applications of this AI.

AI products “made in NRW”

Filter our AI map by “image recognition and understanding”.

AI provider from NRW

Our AI map illustrates who offers AI methods related to image recognition in their portfolio.

Contact us

Dr.-Ing. Stefan Eickeler

Senior Research Engineer

Fraunhofer IAIS
Schloss Birlinghoven
53757 Sankt Augustin

Phone +49 2241 141969

E-Mail senden

Image recognition and understanding

Intelligent document analysis “recognAIze”

With the intelligent document analysis of recognAIze, data from documents can be recognized and evaluated automatically.

Where does the AI application offer the greatest benefit?

The manual review of receipts, invoices and other documents, their digital capture and provision is associated with a high expenditure of time and money in many companies as well as administrative institutions. The solution is provided by intelligent document analysis systems based on optical character recognition (OCR), which, like “recognAIze”, enable fast, simple and automated analysis as well as blind processing of all types of documents. Thanks to artificial intelligence, documents are automatically captured, read, assigned and further processed. Damaged originals, low-quality scans of documents and, in particular, confidential documents are processed without further human intervention and according to high data protection standards.

What are the quality indicators in these types of AI applications?

  • The basis of document analysis is the input data that needs to be analyzed. Since the documents are usually captured in varying image quality, automated image enhancement is very important in the AI application.
  • AI-based optical character recognition (OCR) using artificial neural networks ensures that not only individual text characters are recognized and processed, but also text passages and the structure of a document (e.g. headers or footnotes).
  • Through layout analysis, the AI application can also identify tables within a document and interpret the contents to process invoices automatically in accounting, for example.
  • Particularly in the case of sensitive information, the AI used must be secure and all data must be processed in a DSGVO-compliant manner on German servers or on-premise at the customer’s premises.
  • In the future, handwriting recognition (ICR) will also play a role in the applications in order to open up additional fields of application and achieve a complete transfer of content.

What AI technology powers demonstrator of KI.NRW?

Deep Learning OCR

Optical Character Recognition (OCR) combines Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM), two current approaches in the field of Artificial Intelligence, to develop characters from pixels. It extracts the text from the images and generates a structured XML file for each document with position data of the recognized words and page ranges.

Image enhancement

For the best possible quality of the results, negative influencing factors such as insufficient exposure of the scanned document or curvature or distortion in the image must be equalized. The image enhancement algorithms perform grayscale conversion and binarization for this purpose. In addition, procedures are used to remove curvature and other disturbing factors.

Layout recognition

The layout recognition identifies the structure of text and helps to divide the recognized characters into columns, text sections or headings and to determine a reading order. In this way, table structures can also be recognized and output again as such, e.g. as a csv file. The output format is provided with appropriate metadata.

“Thanks to the methods used for image enhancement, layout and
character recognition, even poor quality documents can be evaluated.”
Dr. Nicolas Flores-Herr
Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS

What does the AI demonstrator show?

The KI.NRW demonstrator “recognAIze” brings AI-supported document analysis to life. Via the application, your own photographed or scanned documents can be uploaded to the system, where they are available for testing the intelligent document analysis. The step by step animations guide you through the AI technologies used in the demonstrator.

Optimize documents

Often, photographed or scanned documents have a fluctuating image quality, and are sometimes bumpy, torn or dirty. Image enhancement processes ensure that even old or damaged documents can be processed. The “recognAIze” demonstrator vividly guides you through the range of optimization options that are essential for high-quality document analysis.

Recognize characters and structures

The accuracy and speed of the OCR engine for intelligent character recognition from “recognAIze” is higher than that of leading market players. Without templates and manual post-processing, the demonstrator recognizes document layouts, e.g. sender information or dates. Even complex text content such as text-around-image elements are reliably recognized by the application.

Understand tables

Tables pose a particular challenge because they can be structured differently from document to document. AI methods are responsible for subdividing table contents according to information types and interpreting the segments individually.

Classify content

The demonstrator “recognAIze” determines the properties of the document, evaluates the individual elements and thereby enables a whole range of subsequent processing. For example, the intelligent classification makes blind processing of confidential documents possible in the first place. This means that information can be aggregated or used pseudonymously without a human being having access to the documents. In this way, sensitive, personal data can be protected better.

Create interfaces

AI-supported document analysis is often at the beginning of a process chain, whether in accounting or in archives. To enable further processing steps, the KI.NRW demonstrator offers various output formats such as XML or PDF.

Are you curious?
Click here to go to the demonstrator!

Where is more information to be found?

Study by KI.NRW

Learn about where we encounter modern language technologies in our everyday and professional lives and the economic opportunities (only available in German language).

AI products “Made in NRW”

Filter our AI map by “language and text comprehension”.

AI provider from NRW

Our AI map illustrates who offers AI methods related to image recognition in their portfolio.

Contact us

Dr. Nicolas Flores-Herr

Business Unit Manager Document Analytics

Fraunhofer IAIS
Schloss Birlinghoven
53757 Sankt Augustin

Phone +49 2241 142532

Email schreiben

Dr. Iuliu Konya

Senior Research Engineer

Fraunhofer IAIS
Schloss Birlinghoven
53757 Sankt Augustin

Phone +49 2241 142543

Email schreiben

 

Marius Nißlmüller

Student assistant Business Development

Fraunhofer IAIS
Schloss Birlinghoven
53757 Sankt Augustin

Email schreiben

To the top