Matej Kandera

Bio: Matej Kandera is an academic researcher. The author has contributed to research in topics: Visual inspection & Clamping. The author has an hindex of 2, co-authored 5 publications receiving 16 citations.

Visual Product Inspection Based on Deep Learning Methods

Abstract: Nowadays, when high industrial productivity is connected with high quality and low product faults, it is common practice to use 100% product quality control. Since the quantities of products are high in mass production and inspection time must be as low as possible, the solution may be to use visual inspection of finished parts via camera systems and subsequent image processing using artificial intelligence. Recently, deep learning has shown itself to be the most appropriate and effective method for this purpose. The present article deals with the above-mentioned method of deep learning, and especially with its application when recognizing certain objects and elements during the visual product inspection.

Design of the clamping system for the CNC machine tool

Abstract: The workpiece clamping and system for automatic workpiece exchange are still topics of great importance. This article describes design of such system implementing interchangeable technological pallets for laboratory training CNC machine tool with confined workspace and limited lengths of feeds. Final design described in article is suitable for implementation in laboratory but also for industrial usage.

Approach to Automated Visual Inspection of Objects Based on Artificial Intelligence

Abstract: The article discusses the possibility of object detector usage in field of automated visual inspection for objects with specific parameters, specifically various types of defects occurring on the surface of a car tire. Due to the insufficient amount of input data, as well as the need to speed up the development process, the Transfer Learning principle was applied in a designed system. In this approach, the already pre-trained convolutional neural network AlexNet was used, subsequently modified in its last three layers, and again trained on a smaller sample of our own data. The detector used in the designed camera inspection system with the above architecture allowed us to achieve the accuracy and versatility needed to detect elements (defects) whose shape, dimensions and location change with each occurrence. The design of a test facility with the application of a 12-megapixel monochrome camera over the rotational table is briefly described, whose task is to ensure optimal conditions during the scanning process. The evaluation of the proposed control system with the quantification of the recognition capabilities in the individual defects is described at the end of the study. The implementation and verification of such an approach together with the proposed methodology of the visual inspection process of car tires to obtain better classification results for six different defect classes can be considered as the main novel feature of the presented research. Subsequent testing of the designed system on a selected batch of sample images (containing all six types of possible defect) proved the functionality of the entire system while the highest values of successful defect detection certainty were achieved from 85.15% to 99.34%.

Modification of Clamping System Designated for Automatic Workpiece Exchange

Abstract: Automatic workpiece exchange is one of the important parts of every flexible manufacturing system. This part largely allows complete automation of the component's production cycle. It also greatly contributes to a significant reduction in the workpiece clamping times and contribute to a reduction in the downtime during the machining process. It also allows eliminating a human factor from the production process, thereby contributing to greater machine utilization and higher machining productivity. This article describes certain modifications of interchangeable pallet system designated for specific training CNC machine tool EMCO Concept Mill 105 situated in the laboratory at the Department of Automation and Production Systems. This machine has considerably limited workspace and length of feeds which are the most limiting parameters during design. Mentioned modifications allow manipulating with designed pallets with the industrial robot or special manipulator instead of previously designed manual handling.

Capabilities of processing and visualization of production facilities diagnostic data

Abstract: One of the main requirements of the safe and sustained workflow of production facilities and machines, is the proper choice and implementation of suitable technical diagnostic method. In the broader sense technical diagnostics includes approaches and methods of technical objects condition determination. This process is based on specific devices parameters inspecting, which values are directly or indirectly caused by certain changes of monitored production facilities condition. In order to reliably detect faults in right time, selected parameters indicating status of the production device should be monitored continuously. At a time when an Internet of Things platforms implemented to the industry business systems is the great temptation from the view of investors and company prestige are these principles common in the process of machine parameters online monitoring and its cross connections within the business realms. This paper describes specific software ThingWorx, which is a great example of sophisticated and universal IoT platform and also describes its application for production facilities diagnostic data collection and visualization.

Proof of Concept

Abstract: The previous part discusses the GI/BSI/DFKI Protection Profile which constitutes after the implementation of the identified improvements as the proposed evaluation methodology for remote electronic voting systems. The result can now be applied to available systems. Currently, there is no system that has been evaluated against the GI/BSI/DFKI Protection Profile or even against the improved version.

Technological Assurance and Features of Fork-Type Parts Machining

Abstract: To provide the machining accuracy of parts on metal-cutting machine-tools fixtures appointed for accurately locating and reliable workpiece clamping are used. The expansion of technological capabilities of modern CNC machine tools necessitates the improvement of design procedures in production planning is needed. The variety of parts and the tendency to increase their complexity, as well as the desire to reduce the cost of time, makes it necessary to find new approaches to the design of tooling. The article proposes the design of a flexible fixture, which provides sufficient tool availability and allows multiaxis machining of fork-type parts at one setup. The ways of intensification and manufacturing process of fork-type parts machining with a significant reduction of auxiliary and preparatory time are offered. Studies performed by numerical simulation methods confirmed that the proposed design meets all the accuracy parameters. The results of static structural, modal, and harmonic analyses confirmed that the proposed fixture had sufficient strength and dynamic stiffness, which allows it to be used with intensive cutting modes that are characteristic of modern machines and cutting tools. The oscillation amplitudes in places of the work surfaces in the proposed device do not exceed the tolerances for the manufacturing of these surfaces.

Approach to Automated Visual Inspection of Objects Based on Artificial Intelligence

Abstract: The article discusses the possibility of object detector usage in field of automated visual inspection for objects with specific parameters, specifically various types of defects occurring on the surface of a car tire. Due to the insufficient amount of input data, as well as the need to speed up the development process, the Transfer Learning principle was applied in a designed system. In this approach, the already pre-trained convolutional neural network AlexNet was used, subsequently modified in its last three layers, and again trained on a smaller sample of our own data. The detector used in the designed camera inspection system with the above architecture allowed us to achieve the accuracy and versatility needed to detect elements (defects) whose shape, dimensions and location change with each occurrence. The design of a test facility with the application of a 12-megapixel monochrome camera over the rotational table is briefly described, whose task is to ensure optimal conditions during the scanning process. The evaluation of the proposed control system with the quantification of the recognition capabilities in the individual defects is described at the end of the study. The implementation and verification of such an approach together with the proposed methodology of the visual inspection process of car tires to obtain better classification results for six different defect classes can be considered as the main novel feature of the presented research. Subsequent testing of the designed system on a selected batch of sample images (containing all six types of possible defect) proved the functionality of the entire system while the highest values of successful defect detection certainty were achieved from 85.15% to 99.34%.

Analysis of the Possibilities of Tire-Defect Inspection Based on Unsupervised Learning and Deep Learning.

Abstract: At present, inspection systems process visual data captured by cameras, with deep learning approaches applied to detect defects. Defect detection results usually have an accuracy higher than 94%. Real-life applications, however, are not very common. In this paper, we describe the development of a tire inspection system for the tire industry. We provide methods for processing tire sidewall data obtained from a camera and a laser sensor. The captured data comprise visual and geometric data characterizing the tire surface, providing a real representation of the captured tire sidewall. We use an unfolding process, that is, a polar transform, to further process the camera-obtained data. The principles and automation of the designed polar transform, based on polynomial regression (i.e., supervised learning), are presented. Based on the data from the laser sensor, the detection of abnormalities is performed using an unsupervised clustering method, followed by the classification of defects using the VGG-16 neural network. The inspection system aims to detect trained and untrained abnormalities, namely defects, as opposed to using only supervised learning methods.

Mobile Applications in Engineering Based on the Technology of Augmented Reality

Abstract: The modern engineering educational environment ensures an extensive use of augmented reality for the successful and deep memorization of graphic material. There are many platforms for implementing augmented reality technology, the most popular of which were analyzed. As a result, the research work is devoted to the visualization of second-order surfaces and engineering objects. The article describes the scenario of augmented reality both with QR codes and authors’ markers. Moreover, the components of the markers and their significance in reproducing the image were described in detail. The proposed methodology includes a step-by-step approach for using the applications based on augmented reality for downloading through the public domain. As a result, the examples of tasks that can be performed using AR were realized. Finally, many markers are proposed in the article, which can be used in the engineering field’s educational process, especially with the mathematical peculiarities.