Depth Data Denoising in Optical Laser Based Sensors for Metal Sheet Flatness Measurement: A Deep Learning Approach

Surface flatness assessment is necessary for quality control of metal sheets manufactured from steel coils by roll leveling and cutting. Mechanical-contact-based flatness sensors are being replaced by modern laser-based optical sensors that deliver accurate and dense reconstruction of metal sheet surfaces for flatness index computation. However, the surface range images captured by these optical sensors are corrupted by very specific kinds of noise due to vibrations caused by mechanical processes like degreasing, cleaning, polishing, shearing, and transporting roll systems. Therefore, high-quality flatness optical measurement systems strongly depend on the quality of image denoising methods applied to extract the true surface height image. This paper presents a deep learning architecture for removing these specific kinds of noise from the range images obtained by a laser based range sensor installed in a rolling and shearing line, in order to allow accurate flatness measurements from the clean range images. The proposed convolutional blind residual denoising network (CBRDNet) is composed of a noise estimation module and a noise removal module implemented by specific adaptation of semantic convolutional neural networks. The CBRDNet is validated on both synthetic and real noisy range image data that exhibit the most critical kinds of noise that arise throughout the metal sheet production process. Real data were obtained from a single laser line triangulation flatness sensor installed in a roll leveling and cut to length line. Computational experiments over both synthetic and real datasets clearly demonstrate that CBRDNet achieves superior performance in comparison to traditional 1D and 2D filtering methods, and state-of-the-art CNN-based denoising techniques. The experimental validation results show a reduction in error than can be up to 15% relative to solutions based on traditional 1D and 2D filtering methods and between 10% and 3% relative to the other deep learning denoising architectures recently reported in the literature.

IMPROVEMENT OF INFORMATIVITY OF COMBINED DIFFERENT-RANGE IMAGES IN MEDICAL DIAGNOSTICS

Задача повышения качества результатов медицинской диагностики и удобства их интерпретации является актуальной на современном этапе развития биомедицинской инженерии. Особый интерес представляют методы визуализации, применимые при диагностике онкологических заболеваний. Повышать достоверность медицинской диагностики таких патологических состояний предлагается посредством совмещения разнодиапазонных изображений, в частности сканов в инфракрасном и видимом диапазонах длин волн. Предлагается методика, в которой два изображения конкретного биообъекта, полученные от датчиков, работающих в разных частотных диапазонах, имеющие одинаковые пространственные параметры и сформированные с общего ракурса, сводятся в общее изображение чересстрочно. Новизна предлагаемой методики заключается в том, что после совмещения изображений производится взаимная передача заданных частей каждого пикселя соседним пикселям по вертикали. В полученном изображении каждый пиксель содержит информацию оптического и инфракрасного изображений в заданных пропорциях. Показано, что предлагаемая методика обеспечивает увеличение информативности в полученном изображении в шесть раз относительно исходных изображений. Предлагаемая методика совмещения разнодиапазонных изображений может быть применена в различных прикладных областях In the medical diagnostics of diseases, it is necessary to obtain the most reliable information in order to obtain the correct diagnosis and, as a result, the correct treatment for the patient. One of the methods of diagnostic studies of oncological diseases of a near-surface nature is to obtain infrared images. It is possible to increase the reliability of information by combining images obtained from thermal imagers, as well as from television video cameras. In this paper, we propose a technique in which two images of a particular object obtained from sensors operating in different frequency ranges, having the same spatial parameters, and formed from the same angle, are interlaced into a common image. The novelty of the proposed method lies in the fact that after combining the images, the specified parts of each pixel are mutually transmitted to the neighboring pixels vertically. In the resulting image, each pixel contains information of optical and infrared images in appropriate proportions. It is shown that the proposed method provides an increase in information content in the resulting image six times relative to the original image. The proposed technique for combining multi-range images can be applied in various areas

A Review on Advances in Automated Plant Disease Detection

Plant diseases cause major yield and economic losses. To detect plant disease at early stages, selecting appropriate techniques is imperative as it affects the cost, diagnosis time, and accuracy. This research gives a comprehensive review of various plant disease detection methods based on the images used and processing algorithms applied. It systematically analyzes various traditional machine learning and deep learning algorithms used for processing visible and spectral range images, and comparatively evaluates the work done in literature in terms of datasets used, various image processing techniques employed, models utilized, and efficiency achieved. The study discusses the benefits and restrictions of each method along with the challenges to be addressed for rapid and accurate plant disease detection. Results show that for plant disease detection, deep learning outperforms traditional machine learning algorithms while visible range images are more widely used compared to spectral images.

Reflective Noise Filtering of Large-Scale Point Cloud Using Multi-Position LiDAR Sensing Data

Signals, such as point clouds captured by light detection and ranging sensors, are often affected by highly reflective objects, including specular opaque and transparent materials, such as glass, mirrors, and polished metal, which produce reflection artifacts, thereby degrading the performance of associated computer vision techniques. In traditional noise filtering methods for point clouds, noise is detected by considering the distribution of the neighboring points. However, noise generated by reflected areas is quite dense and cannot be removed by considering the point distribution. Therefore, this paper proposes a noise removal method to detect dense noise points caused by reflected objects using multi-position sensing data comparison. The proposed method is divided into three steps. First, the point cloud data are converted to range images of depth and reflective intensity. Second, the reflected area is detected using a sliding window on two converted range images. Finally, noise is filtered by comparing it with the neighbor sensor data between the detected reflected areas. Experiment results demonstrate that, unlike conventional methods, the proposed method can better filter dense and large-scale noise caused by reflective objects. In future work, we will attempt to add the RGB image to improve the accuracy of noise detection.