研究目的
To reduce uncertainties in manual ROI selection in SPR imaging systems by proposing and implementing an automatic image segmentation method in an embedded system for real-time applications.
研究成果
The automatic image segmentation method based on watershed algorithm effectively reduces uncertainties in ROI selection for SPR imaging systems, demonstrating good robustness and reproducibility in glucose detection with a linear range of 2.5 to 20.4 mg/mL, R2 of 0.999, and sensitivity of 2.69 a.u./mg/mL. It enables automated data acquisition for various chip designs, improving measurement precision and simplicity.
研究不足
The algorithm may be limited to specific SPR chip designs and patterns; it might not generalize well to all types of surface functionalizations or fluidic channel designs. The embedded system's processing capabilities could constrain real-time performance for very high-resolution images. Noise and image distortions from angle scanning might affect segmentation accuracy if not properly handled.
1：Experimental Design and Method Selection:
The study uses a prism-based intensity modulation image SPR sensor with a Kretschmann configuration. An automatic image segmentation algorithm based on marker-controlled watershed is developed and implemented on a Raspberry Pi 3 embedded system. Methods include histogram slicing, Otsu's method, Canny edge detection, centroid location, histogram equalization, and median filtering for pre-processing and marker assignment.
2：Sample Selection and Data Sources:
SPR images are captured from an experimental setup using Au/Cr thin film chips. Glucose standard solutions with concentrations from 0 to 41.7 mg/mL are used for testing.
3：7 mg/mL are used for testing. List of Experimental Equipment and Materials:
3. List of Experimental Equipment and Materials: Equipment includes a Raspberry Pi 3 embedded system, industrial camera UI-3360CP-NIR-GL, 850 nm infrared laser diode, B270 glass aspheric condenser lens, electron beam evaporation system for Au/Cr deposition, syringe pump, and PMMA chip cover. Materials include borosilicate glass substrates, Cr and Au thin films, PMMA, double-side tape spacer, acetone, ethanol, deionized water, nitrogen gas, and glucose solutions.
4：Experimental Procedures and Operational Workflow:
Prepare Au/Cr chips by cleaning substrates, depositing Cr and Au films, and assembling fluidic chips. Set up the SPR optical system with laser source, prism, and camera. Capture SPR images. Process images using the developed algorithms: pre-processing (contrast enhancement and noise suppression), marker assignment (background extraction, blank region extraction, ROI selection), and watershed segmentation. Perform robustness tests on different chips and glucose detection experiments with triplicate measurements.
5：Data Analysis Methods:
Analyze segmentation results for accuracy and robustness. Calculate re?ective intensity changes, linear range, correlation coefficient, sensitivity, and limit of detection from glucose measurements using statistical methods.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容-
Industrial Camera
UI-3360CP-NIR-GL
IDS
Records SPR images with high frame rate and resolution for processing by the segmentation algorithms.
立即获取同款设备 -
Raspberry Pi 3
3
Raspberry Pi Foundation
Embedded system for executing the image segmentation algorithms in real-time using multi-threading programming.
暂无现货
预约到货通知
-
Infrared Laser Diode
VECSL
Light source to excite surface plasmon resonance in the SPR system.
暂无现货
预约到货通知
-
Aspheric Condenser Lens
Prism coupler in the Kretschmann configuration for SPR excitation.
暂无现货
预约到货通知
-
Electron Beam Evaporation System
Deposits Cr and Au thin films on glass substrates for SPR chips.
暂无现货
预约到货通知
-
Syringe Pump
Pumps glucose solutions into the fluidic chip for SPR measurements.
暂无现货
预约到货通知
-
登录查看剩余4件设备及参数对照表
查看全部
