Advances in Deep Learning for Industrial Inspections
Implementing AI
The Astrocyte software of Teledyne Dalsa is a training tool based on Deep Learning algorithms that include classification, anomaly detection, object detection, segmentation and noise reduction. Non-experts in image processing can now train AI systems while reducing costs significantly.
Inspection of Screws: left: perfect screw; right: bad screw with damaged section encircled. Bottom: heatmap output of a classification algorithm showing the defect Bild: Teledyne Dalsa

The manufacturing industry is being turned on its head as AI, and deep learning transforms the way we create goods and how quality inspection is employed. The combination of software, new deep learning techniques, power of parallel processing, and ease-of-use tools is at the core of this transformation. Traditionally, hundreds, or even thousands, of high-quality, manually classified images were required to train an AI system and create a model that classifies objects with a high degree of predictability. Just gathering this type of dataset has proven to be an obstacle, hindering deep learning adoption into mainstream manufacturing environments. New technology advancements are making it easier for manufacturers to embrace deep learning as part of the inspection process. Today, users train deep learning systems with fewer bad images or even none. While deep learning software for machine vision has been around for more than a decade, it is now becoming more user-friendly and practical. As a result, manufacturers are moving from experimenting with deep learning software to implementing it.

Deep Learning vs. Traditional Methods

Deep learning is ideal for tasks that are difficult to achieve with traditional image processing methods. Typical environments that are suitable for deep learning are those where there is a lot of variables, such as lighting, noise, shape, color, and texture. For example, in food inspection, no two loaves of bread are exactly alike. Each loaf has the same ingredients, and each weighs the same amount, but the shape, color, and texture may be slightly different but still within the range of normality. Another example may be the ripeness of an apple. Ripeness could mean color, softness, or texture; however, there is a range of possibilities where an apple is considered ripe. It is in these types of environments where deep learning shines. Other examples are inspecting the surface finish quality, confirming the presence of multiple items in a kit, detecting foreign objects, and more, to ensure quality throughout the assembly process. A practical example showing the strength of deep learning is scratch inspection on textured surfaces like metal. Some of those scratches are less bright, and their contrast is in the same order of magnitude than that of the textured background itself. Traditional techniques usually fail to locate these types of defects reliably, especially when the shape, brightness, and contrast vary from sample to sample. Figure 1 illustrate scratches inspection on metal sheets. Defects are clearly shown via a heatmap which is a pseudo-color image highlighting the pixels at the location of the defect. Another example of defect inspection is the ability to classify a complex part as being good or bad. For instance metal screws are objects presenting a high degree of variation on their surface make it extremely difficult for traditional algorithms to isolate defects. Deep learning algorithms are very good at inspecting those type of objects.

Simple Classification

Despite the advantage of deep learning over traditional image processing techniques, challenges do exist. First, many users lack the understanding of what is required to achieve success with deep learning. Second, until recently, deep learning required a huge data set to train a system. Many applications have not been able to take advantage of deep learning due to the lack of high quality, manually classified images. In the case where a large data set is available, the next challenge is to label each image. This labeling can be a daunting task because it has to be done by an expert and needs to be error-free. The cases where there is a large number of classes (different groups with a unique label for each) become prone to errors. Subtle labeling errors are one reason for failure to reach satisfactory performance from AI tools. It is painful to realize the amount of wasted time involved before realizing that the failure is due to bad labeling in the original data set. The fact is that a proper data set is the most important item in a particular system and is usually treated as proprietary IP by the user. Typical deep learning applications require hundreds or even thousands of image samples. In more challenging or custom applications, the training model may require up to a million or more image samples. Even if you can get enough images, you have to ensure you have the right mix of good and bad images to meet the parameters of the training model. To achieve the expected results from the training model, you need a balanced data set. This type of training that uses both good and bad examples is called defect detection and is considered a simple classifier. To verify if the training model is accurate, you need to test the model with a new set of images. If the model achieves close to the training set model, it is said that the model generalized well. In the case where the model does poorly on the test set, this tends to reflect that the model remembers all training cases and has not learned what makes an image good or bad and is known as overtraining or overfitting. If the test set does better, then the training set is suspicious (perhaps due to a poor distribution), or the test set is too small. This method is called supervised learning.

Anomaly Detection

Some applications may only have good examples. In many production environments, we see what is acceptable but can never be sure of all possible cases which could cause rejects. There are cases where there is a continuous event of unique new rejects that can occur at very low rates but are still not acceptable. These types of applications could not deploy deep learning effectively due to the lack of bad examples. That is no longer true. New tools have enabled manufacturers to expand the applications that benefit from deep learning. There is a new technique for classification called anomaly detection, where only good examples train a network. In this case, the network recognizes what is considered normal and identifies anything outside that data set as abnormal. If you were to put the good example data set on a graph, it would look like a blob. Anything that falls within the blob classifies as normal, and anything that falls outside the blob classifies as abnormal – an anomaly. Figure 1 and Figure 2 are both solvable with anomaly detection in situations where just a few or even no bad samples are available for training. Anomaly detection tools enable the expansion of deep learning into new applications that could not take advantage of its benefits previously. The inclusion of anomaly detection helps to reduce engineering efforts needed to train a system. If they have the data, non-experts in image processing can train systems while reducing costs significantly. The Astrocyte software of Teledyne Dalsa’s is a training tool based on Deep Learning that includes classification, anomaly detection, object detection, segmentation and noise reduction.

Seiten: 1 2Auf einer Seite lesen

Anzeige

Das könnte Sie auch Interessieren

Bild: Ametek GmbH - Creaform Deutschland GmbH
Bild: Ametek GmbH - Creaform Deutschland GmbH
Kostenfreies Webinar ‚Robot Vision‘

Kostenfreies Webinar ‚Robot Vision‘

Am Dienstag, den 28. September, findet ab 14 Uhr im Rahmen der inVISION TechTalks das kostenlose Webinar ‚Robot Vision‘ statt. In drei 20-minütigen Präsentationen stellen Wenglor, Vecow und Lucid Vision aktuelle Trends bei Software für die Roboterführung, KI-Plattformen für autonome geführte Robotik und Time-of-Flight-Kameras für Robotik und Automation vor.

Bild: Matrix Vision GmbH
Bild: Matrix Vision GmbH
KI-gestützte Sprunganalyse beim Trampolinturnen

KI-gestützte Sprunganalyse beim Trampolinturnen

Wenn es um Podiumsplätze geht, sind Millisekunden und Millimeter entscheidend. Diese Feinheiten sollen für Trampolin-Trainer und -Sportler des Deutschen Turner-Bundes (DTB) zukünftig über ein neues Projekt, welches auf einem bildbasierten System von Simi Reality Motion Systems basiert, visualisiert und dokumentiert werden. Für die passenden Videoaufnahmen sorgen Industriekameras von Matrix Vision.