Marcel Richard Andre Bosling

Text recognition for information retrieval in images of printed circuit boards

In order to achieve an efficient and environment-friendly recycling of printed circuit boards (PCBs), a comprehensive analysis of their material composition is essential. Besides sophisticated chemical and physical methods for a direct material analysis, an indirect method based on information retrieval provides a less costly and more efficient alternative. During the process of information retrieval, PCBs and their components need to be recognized based on their appearance and the corresponding text information. Their material composition is then available through a pre-established database. Therefore, a practical text recognition is necessary for a successful data analysis prior to PCB recycling. Our paper is focusing on two key aspects of text recognition: binarization and final recognition of text objects using optical character recognition (OCR) engines. For binarization of text contents, a novel local thresholding method using an adaptive window size along with background estimation is presented. Several state-of-the-art algorithms and the proposed method were evaluated for comparing their binarization performance on text objects in PCB images. With respect to a data set containing manually created references, our novel method provides superior results. Furthermore, in contrast to previous work on text recognition, an additional evaluation of available open source OCR engines was conducted to asses technical limitations of OCR applications. We show that the quality of text recognition can be significantly improved if the binarization approach accounts for these technical limitations of OCR software. The presented method and results are expected to provide improved OCR performance also in other applications.

Accurate laser triangulation using a perpendicular camera setup to assess the height profile of PCBs

Recycling of printed circuit boards (PCBs) is still subject to current research. For identifying valuable materials in electronical components on the PCBs the height profile of PCBs is a important feature. Laser triangulation with a perpendicular laser is a classical approach to generate height profiles. However, in such a setup it is not possible to acquire different modalities such as panorama images and the height profile simultaneously. The solution is to use a setup in which the laser is slanted and the camera is perpendicular to the PCBs. A new triangulation approach employing raycasting and intersecting a laser plane and a ray of sight is needed to retain accuracy of the height profile generation. Evaluation results show that the accuracy of height measurement with an average error of 0.077 mm is still sufficient for detecting small surface mounted devices on PCBs with a minimal height of about 0.2 mm.

Rotation estimation for printed circuit board recycling

Electronic devices are nowadays an integral part of everyday life. The number of discarded electronic items has grown significantly over the last years. Due to the amount of precious materials used in the manufacturing of these devices recycling of electronic devices is becoming more and more important. Currently, the processes to regain some of these precious materials such as gold, copper, scarce elements etc. are not able to differentiate electronic waste according to its material composition. To enhance these processes, as much information as possible needs to be retrieved per electronic waste item. In particular, information used for the classification of the processed printed circuit boards (PCBs) is important as PCBs are extensively used in electronic devices. One key aspect of this classification process is the estimation of the orientation of the PCBs in this process (e.g. on a conveyor belt). In this paper typical properties of PCBs with respect to orientation estimation are introduced and three different orientation estimation algorithms are evaluated and discussed. These approaches comprise the Hough transform, structure tensors and orientation estimation via Fourier transform. Finally, all presented algorithms are evaluated based on images of PCBs with known orientation. The results indicate, that the Hough transform yields the best results for completely visible PCBs while structure tensors performed best on partially visible PCBs.

Rectification for fronto-parallel camera movement in 3D reconstruction of printed circuit boards

Recycling of printed circuit boards (PCBs) is one of the open challenges in research. For recycling, the identification of the components on PCBs is of great importance. For PCB component classification the height profile of a PCB is one of the important features. One approach to obtain this profile is photogrammetry employing different views of the PCBs surface. Thereby, rectification of the images is essential for highly accurate, dense 3D reconstruction employing block matchers. However, commonly utilized rectification approaches rely on the fundamental matrix which deteriorates in scenes with low depth structure. Thus, a new rectification approach is needed for the case of PCBs. We propose a homography-based approach employing salient image features which are tracked in the images of a camera moved fronto-parallel above the PCB. It can be shown that the accuracy of the 3D reconstruction after applying the proposed rectification allows reconstructing structures at 0.1mm accuracy. This is sufficient for detecting small surface mounted devices with a minimal height of 0.2 mm.

Fingerprinting channel dynamics in indoor low-power wireless networks

Wireless low-power embedded devices are populating indoor environments, where everyday activities drastically impact communication. We explore a statistical approach to identify changes to the communication links state during system operation. The long-term behavior of the link RSSI is modeled with a normal distribution and compared against the model of the most recent measurements. A Welchs t-Test is then employed to identify whether the short-term and long-term link evolutions stem from the same distribution. Upon significant divergence, the long-term model is updated and a significant change in the underlying communication state is inferred. We investigate this technique to efficiently store a compressed fingerprint of the evolution of communication. Considering the memory constraints of low-power embedded systems, this approach allows to gather extensive information on the behavior of communication directly from the deployed network. This fingerprint could then be used to replay the network dynamics in simulation. We implemented the introduced techniques to prove their feasibility. In controlled experiments, we evaluate the reactivity and sensitivity of the approach to changes in the environment, as well as the accuracy of the resulting channel fingerprint.

Can P2P swarm loading improve the robustness of 6LoWPAN data transfer

The recent arrival of 6LoWPAN, an IPv6 variant for low-power wireless devices, allows for the development of IP-based applications for low-power wireless networks like sensor networks. As these networks often suffer from unreliable radio channels or frequent node failure, the question arises how 6LoWPAN-based applications can be hardened against such issues. In this paper, we examine if well-known concepts from the research domain of P2P networks can be applied for increasing the robustness of 6LoWPAN networks. For this purpose, we design and implement a straightforward P2P system for data sharing that is adapted to the low system capabilities of a typical 6LoWPAN device. Our preliminary evaluation shows that P2P mechanisms like swarm loading are able to improve the robustness of a 6LoWPAN application.