Improving the segmentation of digital images by using a modified Otsu’s between-class variance

Singh, Simrandeep; Mittal, Nitin; Singh, Harbinder; Oliva, Diego

doi:10.1007/s11042-023-15129-y

Cited by 7 publications

(3 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Image segmentation is an important element of the system, as it allows for advanced image analysis and understanding. In this work, the denoised image underwent binarization and an opening operation using the OTSU method [31]. Next, the original RGB image was multiplied by the processed binarization matrix to extract the background area.…”

Section: Pre-processing and Segmentationmentioning

confidence: 99%

Automated Grading of Angelica sinensis Using Computer Vision and Machine Learning Techniques

Zhang,

Xiao,

Wang

et al. 2024

Agriculture

View full text Add to dashboard Cite

Angelica sinensis (Oliv.) Diels, a member of the Umbelliferae family, is commonly known as Danggui (Angelica sinensis, AS). AS has the functions of blood tonic, menstrual pain relief, and laxatives. Accurate classification of AS grades is crucial for efficient market management and consumer health. The commonly used method to classify AS grades depends on the evaluator’s observation and experience. However, this method has issues such as unquantifiable parameters and inconsistent identification results among different evaluators, resulting in a relatively chaotic classification of AS in the market. To address these issues, this study introduced a computer vision-based approach to intelligently grade AS. Images of AS at five grades were acquired, denoised, and segmented, followed by extraction of shape, color, and texture features. Thirteen feature parameters were selected based on difference and correlation analysis, including tail area, whole body area, head diameter, G average, B average, R variances, G variances, B variances, R skewness, G skewness, B skewness, S average, and V average, which exhibited significant differences and correlated with grades. These parameters were then used to train and test both the traditional back propagation neural network (BPNN) and the BPNN model improved with a growing optimizer (GOBPNN). Results showed that the GOBPNN model achieved significantly higher average testing precision, recall, F-score, and accuracy (97.1%, 95.9%, 96.5%, and 95.0%, respectively) compared to the BPNN model. The method combining machine vision technology with GOBPNN enabled efficient, objective, rapid, non-destructive, and cost effective AS grading.

show abstract

Section: Pre-processing and Segmentationmentioning

confidence: 99%

Automated Grading of Angelica sinensis Using Computer Vision and Machine Learning Techniques

Zhang,

Xiao,

Wang

et al. 2024

Agriculture

View full text Add to dashboard Cite

show abstract

“…The Otsu method is mostly used in binary segmentation, but multi-level Otsu thresholding can be the solution when more than two classes are required. With the input image having a multimodal histogram, more than one threshold is necessary to segment the image properly [19]. So, the Otsu technique is used to find the threshold values in multi-Otsu thresholding.…”

Section: Step 2: Segmentation Of Rgb Imagementioning

confidence: 99%

GeoBIA-based semi-automated landslide detection using UAS data: a case study of Uttarakhand Himalayas

Kundal

Chowdhury

Bhardwaj

et al. 2023

SPIE Future Sensing Technologies 2023

View full text Add to dashboard Cite

Recent developments in the field of Geographic Object-Based Image Analysis (GeoBIA) have been utilized for the Automation of Landslide detection. In this study, we have tried to develop a semi-automated detection methodology applying concepts of GeoBIA on High-Resolution Earth Observation imagery acquired from an Uncrewed Aerial Systems also referred to as Uncrewed Aerial Vehicles (UAV). The study area is in the Himalayan state of Uttarakhand, India. The UAV was flown over a landslide site. The UAV data was processed for deriving photogrammetric products (Digital Elevation Model and Orthomosaic). The methodology implements the segmentation and classification of UAV images using Multi-otsu thresholding method and machine learning algorithm of random forest. It incorporates spectral (RGB), textural (GLCM entropy and GLCM angular second moment), morphological (sky view factor), and topographical (elevation, slope, curvature) features derived from UAV photogrammetric products. When determining landslide locations is of utmost concern, this system's ability to detect landslides quickly and effectively gives it a viable alternative to manual procedures for landslide mapping across wide areas. The developed system was able to detect 86% of the total landslide area.

show abstract

“…These methods can be used to immediately determine the threshold, but they suffer the drawbacks of a time-consuming process and significant human error. Theoretical methods, which can be categorized as global or local methods, are based on certain threshold calculation theories [29], such as the Otsu algorithm [30,31], iteration algorithm [32], image histograms [33], entropy criterion [34], gray wolf optimizer [35], and edge detection [36]. Although these theories could facilitate threshold calculation, they are not all appropriate for clay SEM image processing due to the very small size of soil particles and pores and the difficulty in identifying contours.…”

Section: Introductionmentioning

confidence: 99%

Determining Effective Threshold Range of Image Pixel Values for Municipal Waste-Contaminated Clay

Zhang,

Cao,

Guo

2024

Applied Sciences

View full text Add to dashboard Cite

Currently, more attention has been given to the study of the microstructure of contaminated clay, while few studies have focused on the analysis of SEM images of clay from a threshold perspective. The purpose of this paper was to determine the threshold range of image pixel values for municipal waste-contaminated clay by observing and processing SEM images of municipal waste-contaminated clay under different seepage depths and concentrations. In this study, municipal waste-contaminated clay samples were obtained at different depths in a column test exhibiting different CaCO3 and CH3COOH concentrations, imaged through SEM and subjected to binarization. Three steps for determining the pixel threshold value of the SEM images were described, including preprocessing and binarizing the SEM images in Image Pro Plus 6.0 (IPP) and morphological processing by the Pores (Particles) and Cracks Analysis System V2.3 (PCAS). The apparent void ratio of contaminated clay and the fractal dimension of clay particles were applied as evaluation parameters. The determined effective threshold range of the image pixel values of municipal waste-contaminated clay was 110–140. The accuracy of this method was verified in three aspects: the relationship between the threshold value and the apparent void ratio, the variation in pore blockage with seepage depth, and the characteristics of the compaction factor with seepage depth and pollutant concentration.

show abstract

Improving the segmentation of digital images by using a modified Otsu’s between-class variance

Cited by 7 publications

References 48 publications

Automated Grading of Angelica sinensis Using Computer Vision and Machine Learning Techniques

Automated Grading of Angelica sinensis Using Computer Vision and Machine Learning Techniques

GeoBIA-based semi-automated landslide detection using UAS data: a case study of Uttarakhand Himalayas

Determining Effective Threshold Range of Image Pixel Values for Municipal Waste-Contaminated Clay

Contact Info

Product

Resources

About