Search Results (98)

Computer vision and image processing techniques have had great success in the food and drink industry. These technologies are used to analyse images, convert images to greyscale, and extract high-dimensional numerical data from the images; however, when it comes to real-time grain and rice milling processes, this technology has several limitations compared to other applications. Currently, milled rice image samples are collected and separated to avoid one contacting the another during analysis. This approach is not suitable for real-time industrial implementation. However, real-time analysis can be accomplished by utilising artificial intelligence (AI) and machine learning (ML) approaches instead of traditional quality assessment methods, such as manual inspection, which are labour-intensive, time-consuming, and prone to human error. To address these challenges, this paper presents a novel approach for real-time rice morphology analysis during milling by integrating You Only Look Once version 8 (YOLOv8) instance segmentation and Oriented Bounding Box (OBB) detection models. While instance segmentation excels in detecting and classifying both touching and overlapping grains, it underperforms in precise size estimation. Conversely, the object-oriented bounding box detection model provides more accurate size measurements but struggles with touching and overlapping grains. Experiments demonstrate that the hybrid system resolves key limitations of standalone models: instance segmentation alone achieves high detection accuracy (92% mAP@0.5) but struggles with size errors (0.35 mm MAE), while OBB alone reduces the size error to 0.12 mm MAE but falters with complex grain arrangements (88% mAP@0.5). By combining these approaches, our unified pipeline achieves superior performance, improving detection precision (99.5% mAP@0.5), segmentation quality (86% mask IoU), and size estimation (0.10 mm MAE). This represents a 71% reduction in size error compared to segmentation-only models and a 6% boost in detection accuracy over OBB-only methods. This study highlights the potential of advanced deep learning techniques in enhancing the automation and optimisation of quality control in rice milling processes. Full article

Process control and quality assurance of wire + arc additive manufacturing (WAAM) and automated welding rely heavily on in-process monitoring videos to quantify variables such as melt pool geometry, location and size of droplet transfer, arc characteristics, etc. To enable feedback control based upon this information, an automatic and robust segmentation method for monitoring of videos and images is required. However, video segmentation in WAAM and welding is challenging due to constantly fluctuating arc brightness, which varies with deposition and welding configurations. Additionally, conventional computer vision algorithms based on greyscale value and gradient lack flexibility and robustness in this scenario. Deep learning offers a promising approach to WAAM video segmentation; however, the prohibitive time and cost associated with creating a well-labelled, suitably sized dataset have hindered its widespread adoption. The emergence of large computer vision models, however, has provided new solutions. In this study a semi-automatic annotation tool for WAAM videos was developed based upon the computer vision foundation model SAM and the video object tracking model XMem. The tool can enable annotation of the video frames hundreds of times faster than traditional manual annotation methods, thus making it possible to achieve rapid quantitative analysis of WAAM and welding videos with minimal user intervention. To demonstrate the effectiveness of the tool, three cases are demonstrated: online wire position closed-loop control, droplet transfer behaviour analysis, and assembling a dataset for dedicated deep learning segmentation models. This work provides a broader perspective on how to exploit large models in WAAM and weld deposits. Full article

This paper presents a novel forensic watermarking method for digital evidence distribution in non-cloud environments. The approach addresses the critical need for the secure sharing of Joint Photographic Experts Group (JPEG) images in forensic investigations. The method utilises an adaptive Discrete Cosine Transform–Discrete Wavelet Transform (DCT-DWT) domain technique to embed a 64-bit watermark in both stand-alone JPEGs and those within forensic disk images. This occurs without alterations to disk structure or complications to the chain of custody. The system implements uniform secure randomisation and recipient-specific watermarks to balance security with forensic workflow efficiency. This work presents the first implementation of forensic watermarking at the disk image level that preserves structural integrity and enables precise leak source attribution. It addresses a critical gap in secure evidence distribution methodologies. The evaluation occurred on extensive datasets: 1124 JPEGs in a forensic disk image, 10,000 each of BOSSBase 256 × 256 and 512 × 512 greyscale images, and 10,000 COCO2017 coloured images. The results demonstrate high imperceptibility with average Peak Signal-to-Noise Ratio (PSNR) values ranging from 46.13 dB to 49.37 dB across datasets. The method exhibits robust performance against geometric attacks with perfect watermark recovery (Bit Error Rate (BER) = 0) for rotations up to 90° and scaling factors between 0.6 and 1.5. The approach maintains compatibility with forensic tools like Forensic Toolkit FTK and Autopsy. It performs effectively under attacks including JPEG compression (QF ≥ 60), filtering, and noise addition. The technique achieves high feature match ratios between 0.684 and 0.690 for a threshold of 0.70, with efficient processing times (embedding: 0.0347 s to 0.1187 s; extraction: 0.0077 s to 0.0366 s). This watermarking technique improves forensic investigation processes, particularly those that involve sensitive JPEG files. It supports leak source attribution, preserves evidence integrity, and provides traceability throughout forensic procedures. Full article

Accurate localisation is key for the autonomy of mobile robots. Fiducial localisation utilises relative positions of markers physically deployed across an environment to determine a localisation estimate for a robot. Fiducial markers are strictly designed, with very limited flexibility in appearance. This often results in a “trade-off” between visual customisation, library size, and occlusion resilience. Many fiducial localisation approaches vary in their position estimation over time, leading to instability. The Stable Fiducial Marker System (STag) was designed to address this limitation with the use of a two-stage homography detection. Through its combined square and circle detection phases, it can refine detection stability. In this work, we explore the utility of STag as a basis for a stable mobile robot localisation system. Key marker restrictions are addressed in this work through contributions of three new chromatic STag marker types. The hue/greyscale STag marker set addresses constraints in customisability, the high-capacity STag marker set addresses limitations in library size, and the high-occlusion STag marker set improves resilience to occlusions. These are designed with compatibility with the STag detection system, requiring only preprocessing steps for enhanced detection. They are assessed against the existing STag markers and each shows clear improvements. Further, we explore the viability of various materials for marker fabrication, for use in outdoor and low-light conditions. This includes the exploration of “active” materials which induce effects such as retro-reflectance and photo-luminescence. Detection rates are experimentally assessed across lighting conditions, with “active” markers assessed on the practicality of their effects. To encapsulate this work, we have developed a full end-to-end deployment for fiducial localisation under the STag system. It is shown to function for both on-board and off-board localisation, with deployment in practical robot trials. As a part of this contribution, the associated software for marker set generation/detection, physical marker fabrication, and end-to-end localisation has been released as an open source distribution. Full article

The use of artificial intelligence-based techniques to solve engineering problems is increasing. One of the most challenging tasks facing industry is the timely diagnosis of failures in electromechanical systems, as they are an essential part of production systems. In this sense, the earlier the detection, the higher the economic loss reduction. For this reason, this work proposes the development of a new methodology based on infrared thermography and an artificial intelligence-based classifier for the detection of multiple faults in an electromechanical system. The proposal combines the intensity profile of the grey-scale image, the use of Fast Fourier Transform and an artificial neural network to perform the detection of twelve states for the state of an electromechanical system: healthy, bearing defect, broken rotor bar, misalignment and gear wear on the gearbox. From the experimental setup, 50 thermographic images were obtained for each state. The method was implemented and tested under different conditions to verify its reliability. The results show that the precision, accuracy, recall and F1-score are higher than 99%. Thus, it can be concluded that it is possible to detect multiple conditions in an electromechanical system using the intensity profile and an artificial neural network, achieving good accuracy and reliability. Full article

In this work, an analytical method in the frequency domain is proposed for the design of two-dimensional digital FIR differentiators. This technique uses an approximation based on two methods: the Chebyshev series and the Fourier series, which, finally, lead to a trigonometric polynomial, which is a remarkably precise approximation of the transfer function of the ideal differentiator. The digital differentiator is applied to three test images, one greyscale image and two binary images, and simulation results show its performance in the processing task. Also, based on the fact that this 2D differentiator is separable on the two frequency axes, we propose an efficient implementation at the system level, using polyphase filtering. The designed digital differentiator is very accurate and efficient, having a high level of parallelism and reduced computational complexity. Full article

The sustainable development of agriculture is closely related to the adoption of precision agriculture techniques, and accurate crop classification is a fundamental aspect of this approach. This study explores the application of machine learning techniques to crop classification by integrating RGB images and multispectral data acquired by UAVs. The study focused on five crops: rice, soybean, red bean, wheat, and corn. To improve classification accuracy, the researchers extracted three key feature sets: band values and vegetation indices, texture features extracted from a grey-scale co-occurrence matrix, and shape features. These features were combined with five machine learning models: random forest (RF), support vector machine (SVM), k-nearest neighbour (KNN) based, classification and regression tree (CART) and artificial neural network (ANN). The results show that the Random Forest model consistently outperforms the other models, with an overall accuracy (OA) of over 97% and a significantly higher Kappa coefficient. Fusion of RGB images and multispectral data improved the accuracy by 1–4% compared to using a single data source. Our feature importance analysis showed that band values and vegetation indices had the greatest impact on classification results. This study provides a comprehensive analysis from feature extraction to model evaluation, identifying the optimal combination of features to improve crop classification and providing valuable insights for advancing precision agriculture through data fusion and machine learning techniques. Full article

This video article summarizes a case study involving the use of pelvic sentinel lymph node (SLN) biopsy for endometrial cancer (EC) staging and treatment utilizing a multi-modal infrared signal technology. This innovative approach combines cervical injection of fluorescent dye indocyanine green (ICG) and near-infrared imaging to enhance SLN detection rates in early-stage EC patients. The study showcases the successful application of advanced technology in improving surgical staging procedures and reducing postoperative morbidity for patients. Multi-modal infrared signal technology consists of different modes of fluorescence imaging used to identify lymph nodes based on near-infrared signals. Each mode serves a specific purpose: overlay image combines white light and near-infrared signals in green, monochromatic visualization shows near-infrared signal in greyscale, and intensity map combines signals in a color scale to differentiate signal intensity. Yellow denotes strong near-infrared signals while blue represents weaker signals. By utilizing a multi-modal approach, surgeons can accurately identify and remove SLN, thus avoiding unnecessary removal of secondary or tertiary echelons. Full article

Rheumatoid Arthritis (RA) is a chronic autoimmune illness that occurs in the joints, resulting in inflammation, pain, and stiffness. X-ray examination is one of the most common diagnostic procedures for RA, but manual X-ray image analysis has limitations because it is a time-consuming procedure and is prone to errors. A specific algorithm aims to a lay stable and accurate segmenting of carpal bones from hand bone images, which is vitally important for identifying rheumatoid arthritis. The algorithm demonstrates several stages, starting with Carpal bone Region of Interest (CROI) specification, dynamic thresholding, and Gray Level Co-occurrence Matrix (GLCM) application for texture analysis. To get the clear edges of the image, the component is first converted to the greyscale function and thresholding is carried out to separate the hand from the background. The pad region is identified to obtain the contours of it, and the CROI is defined by the bounding box of the largest contour. The threshold value used in the CROI method is given a dynamic feature that can separate the carpal bones from the surrounding tissue. Then the GLCM texture analysis is carried out, calculating the number of pixel neighbors, with the specific intensity and neighbor relations of the pixels. The resulting feature matrix is then employed to extract features such as contrast and energy, which are later used to categorize the images of the affected carpal bone into inflamed and normal. The proposed technique is tested on a rheumatoid arthritis image dataset, and the results show its contribution to diagnosis of the disease. The algorithm efficiently divides carpal bones and extracts the signature parameters that are critical for correct classification of the inflammation in the cartilage images. Full article

In the study of aircraft wake target detection, as the wake evolves and develops, the detection area of the LiDAR often shows the presence of two distinct vortices, one on each side. Sometimes, only a single wake vortex may be present. This can lead to a reduction in the accuracy of wake detection and an increased likelihood of missed detections, which may have a significant impact on the flight safety. Hence, we propose an algorithm based on the YOLOv8n–CBAM–EfficientNetV2 model for wake detection. The algorithm incorporates the lightweight network of EfficientNetV2 and the Convolutional Block Attention Module (CBAM) based on the YOLOv8n model, which achieves the lightweight improvement in the YOLOv8n algorithm and the improvement in detection accuracy. First, this study classifies the wake vortices in the wake greyscale images obtained at Hong Kong International Airport, based on the Range–Height Indicator (RHI) scanning characteristics of the LiDAR and the symmetry of the wake vortex pairs. The classification is used to detect left and right vortices for more accurate wake detection in wind field images, which thereby improves the precision rate of target detection. Subsequently, experiments are conducted using a YOLOv8n–CBAM–EfficientNetV2 model for aircraft wake detection. Finally, the performance of the YOLOv8n–CBAM–EfficientNetV2 model is analysed. The results show that the algorithm proposed in this study can achieve a 96.35% precision rate, 93.58% recall rate, 95.06% F1-score, and 250 frames/s. The results show that the method proposed in this study can be effectively applied in aircraft wake detection. Full article

This paper presents the design of a 40 × 40 micro-light-emitting (micro-LED) test array based on a 20 mm × 20 mm substrate. A study of the relationship between luminous brightness, driving current, and driving voltage revealed that the data voltages of the red, green, and blue micro-LED array from 0 to 255 grey levels are 0.31 V, 0.29 V, and 0.30 V, respectively, under the condition that the target brightness of the white field is 1000 nits and the color temperature is 9300 K. The brightness range of the red micro-LED array is 64.8–101.2%, the brightness range of the green micro-LED array is 66.5–102.8%, and the brightness range of the blue micro-LED array is 53.5–129.2%. In order to overcome the luminance nonuniformity, a grey level depth of 12 -bit is required. A 10T3C pixel driver circuit based on low-temperature polysilicon (LTPS) with a depth of 12 bits greyscale is designed and fabricated into the micro-LED display. A brightness uniformity of 84.1–97.1% can be achieved by brightness correction combined with a 12-bit greyscale depth system for micro-LED display. This provides a valuable reference point for subsequent improvements in the quality of micro-LED displays. Full article

Inclusive design does not stop at removing physical obstacles such as staircases. It also involves identifying architectural features that impose sensory burdens, such as repetitive visual patterns that are known to potentially cause dizziness or visual discomfort. In order to assess their influence on human gait and its stability, three repetitive patterns—random dots, repetitive stripes, and repetitive waves (Lisbon pattern)—were displayed in a coloured and greyscale variant in a virtual reality (VR) environment. The movements of eight participants were recorded using a motion capture system and electromyography (EMG). During all test conditions, a significant increase in the muscular activity of leg flexor muscles was identified just before touchdown. Further, an increase in the activity of laterally stabilising muscles during the swing phase was observed for all of the test conditions. The lateral and vertical centre of mass (CoM) deviation was statistically evaluated using a linear mixed model (LMM). The patterns did cause a significant increase in the CoM excursion in the vertical direction but not in the lateral direction. These findings are indicative of an inhibited and more cautious gait style and a change in control strategy. Furthermore, we quantified the induced discomfort by using both algorithmic estimates and self-reports. The Fourier-based methods favoured the greyscaled random dots over repetitive stripes. The colour metric favoured the striped pattern over the random dots. The participants reported that the wavey Lisbon pattern was the most disruptive. For architectural and structural design, this study indicates (1) that highly repetitive patterns should be used with care in consideration of their impact on the human visuomotor system and its behavioural effects and (2) that coloured patterns should be used with greater caution than greyscale patterns. Full article

This study presents a method to enhance the contrast and luminosity of fundus images with boundary reflection. In this work, 100 retina images taken from online databases are utilized to test the performance of the proposed method. First, the red, green and blue channels are read and stored in separate arrays. Then, the area of the eye also called the region of interest (ROI) is located by thresholding. Next, the ratios of R to G and B to G at every pixel in the ROI are calculated and stored along with copies of the R, G and B channels. Then, the RGB channels are subjected to average filtering using a 3 × 3 mask to smoothen the RGB values of pixels, especially along the border of the ROI. In the background brightness estimation stage, the ROI of the three channels is filtered by binomial filters (BFs). This step creates a background brightness (BB) surface of the eye region by levelling the foreground objects like blood vessels, fundi, optic discs and blood spots, thus allowing the estimation of the background illumination. In the next stage, using the BB, the luminosity of the ROI is equalized so that all pixels will have the same background brightness. This is followed by a contrast adjustment of the ROI using CLAHE. Afterward, details of the adjusted green channel are enhanced using information from the adjusted red and blue channels. In the color correction stage, the intensities of pixels in the red and blue channels are adjusted according to their original ratios to the green channel before the three channels are reunited. The resulting color image resembles the original one in color distribution and tone but shows marked improvement in luminosity and contrast. The effectiveness of the approach is tested on the test images and enhancement is noticeable visually and quantitatively in greyscale and color. On average, this method manages to increase the contrast and luminosity of the images. The proposed method was implemented using MATLAB R2021b on an AMD 5900HS processor and the average execution time was less than 10 s. The performance of the filter is compared to those of two other filters and it shows better results. This technique can be a useful tool for ophthalmologists who perform diagnoses on the eyes of diabetic patients. Full article

In modern digital photogrammetry, an image is usually registered via a digital matrix with an array of colour filters. From the registration of the image until feature points are detected on the image, the image is subjected to a series of calculations, i.e., demosaicing and conversion to greyscale, among others. These algorithms respond differently to the varying light spectrum of the scene, which consequently results in the feature location changing. In this study, the effect of scene illumination on the localisation of a feature in an image is presented. The demosaicing and greyscale conversion algorithms that produce the largest and smallest deviation of the feature from the reference point were assessed. Twelve different illumination settings from polychromatic light to monochromatic light were developed and performed, and five different demosaicing algorithms and five different methods of converting a colour image to greyscale were analysed. A total of 300 different cases were examined. As the study shows, the lowest deviation in the polychromatic light domain was achieved for light with a colour temperature of 5600 K and 5000 K, while in the monochromatic light domain, it was achieved for light with a green colour. Demosaicing methods have a significant effect on the localisation of a feature, and so the smallest feature deviation was achieved for smooth hue-type demosaicing, while for greyscale conversion, it was achieved for the mean type. Demosaicing and greyscale conversion methods for monochrome light had no effect. The article discusses the problem and concludes with recommendations and suggestions in the area of illuminating the scene with artificial light and the application of the algorithms, in order to achieve the highest accuracy using photogrammetric methods. Full article

The vibration signals collected by acceleration sensors are interspersed with noise interference, which increases the difficulty of fault diagnosis for rolling bearings. For this reason, a rolling bearing fault diagnosis method based on complete ensemble empirical model decomposition with adaptive noise (CEEMDAN) and improved convolutional neural network (CNN) is proposed. Firstly, the original vibration signal is decomposed into a series of intrinsic modal function (IMF) components using the CEEMDAN algorithm, the components are filtered according to the correlation coefficients and the signals are reconstructed. Secondly, the reconstructed signals are converted into a two-dimensional grey-scale map and input into a convolutional neural network to extract the features. Lastly, the features are inputted into a support vector machine (SVM) with the optimised parameters of the grey wolf optimiser (GWO) to perform the identification and classification. The experimental results show that the rolling bearing fault diagnosis method based on CEEMDAN and CNN-SVM proposed in this paper can significantly reduce the noise interference, and its average fault diagnosis accuracy is as high as 99.25%. Therefore, it is feasible to apply it in the field of rolling bearing fault diagnosis. Full article