Instant testing and non-contact diagnosis for photovoltaic cells using K-means clustering and associated hyperspectral imaging

SN Applied Sciences, Journal Year: 2023, Volume and Issue: 5(8)

Published: July 12, 2023

Abstract Renewable energy, particularly solar has experienced remarkable growth in recent years. However, the integrity of photovoltaic (PV) cells can degrade over time, necessitating non-destructive testing and evaluation (NDT-NDE) for quality control during production in-service inspection. Hyperspectral (HS) imaging emerged as a promising technique defect identification PV based on their spectral signatures. This study utilizes HS imager to establish diffuse reflectance spectra signature two groups cells: working non-working. A non-contact photoluminescence imaging-based methodology is employed, using halogen lamp an illumination source replicate sunlight. Our findings reveal that non-working regions be differentiated from within 400–600 nm wavelength range, with optimal candidate peak frequency 450 nm. To accurately group active constructed images at nm, we employ image processing strategy combines K-means clustering (K-mc) contour delineation. Specifically, K-mc K = 8 used efficiently precisely regions. We demonstrate effectiveness this proposed approach compare it traditional infrared (IR) techniques. implemented conventional camera filter NDT-NDE exterior-mounted panels. Overall, technique, coupled K-mc, offers rapid effective means identifying defects cells, outperforming IR advancement contributes increased efficiency extended lifespan

Language: Английский

---

Instant plastic waste detection on shores using laser-induced fluorescence and associated hyperspectral imaging

Optical and Quantum Electronics, Journal Year: 2024, Volume and Issue: 56(5)

Published: March 25, 2024

Abstract Plastic pollution is a rising environmental issue, with millions of tons plastic debris collecting in the world's seas and on its shores. Hyperspectral imaging (HSI) has become increasingly widely used as more precise approach that can identify targets remote sensing aquatic missions. The interference from other beach materials, need for proper identification litter types make identifying dumped plastics sand-surrounded beaches challenging. This study lays groundwork physical laboratory setting images captured by hyperspectral (HS) imager. suggested testing setup included development fluorescence signature target theater operations (low-density polyethylene (LD-PE) wood surrounded sand) detecting polymers simulated environment using laser-induced (LIF) approach. Initially broadband-spectrum light, strong sample diffuse reflectance contrast observed at wavelengths between 400 460 nm. Next, dedicated LIF system discovery was developed an ultraviolet (UV) laser source. Initial findings show there distinct signal 450 nm 750 wood. Our pilot studies support current efforts to determine optimum spectral these will appear clarity shorelines inexpensive imagery combined our UV approach, which may have impact applications detection pollution. knowledge gained this be construct reliable aerial conventional cameras waste monitoring management.

Language: Английский

---

Multi-wavelength interference phase imaging for automatic breast cancer detection and delineation using diffuse reflection imaging

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Jan. 3, 2024

Abstract Millions of women globally are impacted by the major health problem breast cancer (BC). Early detection BC is critical for successful treatment and improved survival rates. In this study, we provide a progressive approach using multi-wavelength interference (MWI) phase imaging based on diffuse reflection hyperspectral (HS) imaging. The proposed findings measurement pattern between blue (446.6 nm) red (632 wavelengths. We consider implementing comprehensive image processing categorization method use Fast Fourier (FF) transform analysis pertaining to change in refractive index tumor normal tissue. observed that growth affects tissue organization dramatically, as seen persistently increased variance tumors compared areas. Both malignant had different depth data collected from it was analyzed. To enhance ex-vivo tissue, developed validated training classifier algorithm specifically designed categorizing HS cube data. Following application signal normalization with FF algorithm, our methodology achieved high level performance specificity ( Spec ) 94% sensitivity Sen 90.9% 632 nm acquired categorization, preliminary specimens under investigation. Notably, successfully leveraged unstained samples create 3D phase-resolved images effectively highlight distinctions reflectance features cancerous healthy Preliminary revealed might be able assist specialists safely excising areas assessing bed following resection automatically at depths. This investigation result an effective "in-vivo" disease description utilizing optical technology typical RGB camera wavelength-specific operation quantitative MWI methodology.

Language: Английский

---

Correlation analysis between texture features and elasticity of skin hyperspectral images in the near‐infrared band

Skin Research and Technology, Journal Year: 2024, Volume and Issue: 30(3)

Published: March 1, 2024

Abstract Background/purpose Skin elasticity was used to evaluate healthy and diseased skin. Correlation analysis between image texture characteristics skin performed study the feasibility of assessing using a non‐contact method. Materials methods images in near‐infrared band were acquired hyperspectral camera, obtained elastimeter. Texture features mean, standard deviation, entropy, contrast, correlation, homogeneity, energy extracted from images, correlation with performed. Results The features, had highest on side eye under arm, mean suitable for distinguishing according body part. Conclusion In this study, we analyses various parts band, confirming significant some parts. It is expected that will be as cornerstone evaluation research methods.

Language: Английский

---

Instant identification of dental white spot using K-means algorithm via laser-induced fluorescence and associated hyperspectral imaging

Journal of Photochemistry and Photobiology B Biology, Journal Year: 2023, Volume and Issue: 245, P. 112749 - 112749

Published: June 21, 2023

Language: Английский

---

Machine Learning-Based Classification of Skin Cancer Hyperspectral Images

Procedia Computer Science, Journal Year: 2023, Volume and Issue: 225, P. 2856 - 2865

Published: Jan. 1, 2023

Among the different contactless techniques for medical diagnosis, hyperspectral imaging has gained relevance due to high accuracy in tissues classification. Several have been proposed elaborate these images, ranging from traditional machine learning methods deep algorithms. This paper evaluates three popular methods, namely Support Vector Machine (SVM), Random Forest (RF) and eXtreme Gradient Boosting (XGBoost) considering a dataset of skin cancer images. The study demonstrates that algorithms are suitable data classification, particularly when small dataset.

Language: Английский

---

Enhancing automatic inspection and characterization of carbon fiber composites through hyperspectral diffuse reflection analysis and k-means clustering

The International Journal of Advanced Manufacturing Technology, Journal Year: 2024, Volume and Issue: 131(9-10), P. 4935 - 4951

Published: Feb. 29, 2024

Abstract Numerous industries utilize carbon fiber composites (CFC) for their exceptional strength-to-weight ratio and stiffness. However, inherent manufacturing defects such as voids delamination can undermine the material’s structural integrity performance. This study introduces an advanced imaging technique employing hyperspectral (HSI) to effectively detect characterize flaws within CFC materials. HSI provides high-resolution spectral data, enabling precise analysis of material properties. Initial observations indicate a distinctive peak variability in dispersed reflectance spanning 440 nm 600 both pristine surfaces defective regions sheets. The proposed methodology entails automated characterization through combination clustering technique, k-means (k-mc). application k-mc facilitates rapid accurate categorization defect locations. Statistical reveals mean standard deviation (SD) values 0.34 0.33, respectively, normal surfaces, compared 0.43 0.29 cracked sections. These discernible variations enable differentiation between defect-free specimens. By signatures at 445 nm, 546 585 optical markers, accurately measures fracture penetration depths. Implementation this approach generates 3D-resolved images, offering comprehensive visualization imperfections. presents more objective strategy identification categorization. unique holds significant potential industrial applications, particularly scenarios necessitating efficient evaluation numerous components.

Language: Английский

---

Delineation and detection of breast cancer using novel label-free fluorescence

BMC Medical Imaging, Journal Year: 2023, Volume and Issue: 23(1)

Published: Sept. 16, 2023

Accurate diagnosis of breast cancer (BC) plays a crucial role in clinical pathology analysis and ensuring precise surgical margins to prevent recurrence.

Language: Английский

---

Visualization of Subcutaneous Blood Vessels Based on Hyperspectral Imaging and Three-Wavelength Index Images

Sensors, Journal Year: 2023, Volume and Issue: 23(21), P. 8895 - 8895

Published: Nov. 1, 2023

Blood vessel visualization technology allows nursing staff to transition from traditional palpation or touch locate the subcutaneous blood vessels visualized localization by providing a clear visual aid for performing various medical procedures accurately and efficiently involving vessels; this can further improve first-attempt puncture success rate reduce pain of patients. We propose novel technique hyperspectral in human skin. An experiment with six participants different skin types, race, nationality backgrounds is described. A mere separation spectral layers types shown be insufficient. The use three-wavelength indices imaging has significant improvement quality results compared using only two-wavelength indices. This attributed an increase contrast ratio, which as high 25%. implement finding new index formulae based on exhaustive search binary blood-vessel image obtained through expert assessment. As result search, formula was deduced, allowing high-contrast images generated any type.

Language: Английский

---

Simplifying vein detection for intravenous procedures: A comparative assessment through near‐infrared imaging system

International Journal of Imaging Systems and Technology, Journal Year: 2024, Volume and Issue: 34(3)

Published: March 29, 2024

Abstract The intravenous (IV) injection procedure can be a challenging task, especially for individuals with thin veins, obesity, or patients damaged and pigmented skin. Therefore, the IV necessitates portable medical device that used academic demonstrations to train students by health care professionals perform venipuncture. Vein visualization vein detector is principally based on interaction of blood components wavelengths electromagnetic spectrum (EMS). In this paper, we first present process image formation in spectral band near‐infrared region (NIR) EMS. Then, introduce acquisition system processing extract veins as noninvasive detection method. A Raspberry Pi (Model 4B), along night vision camera, serves an tool capture skin area illuminated NIR. Following this, data transferred laptop where it filtered processed using Python tools before being viewed monitor. results achieved through are quite encouraging, recognition between adjacent tissues from sample clearly marked. functionality, accuracy, simplicity associated make potential placement morphological study disease detection.

Language: Английский

---