threshold segmentation
Recently Published Documents


TOTAL DOCUMENTS

269
(FIVE YEARS 86)

H-INDEX

11
(FIVE YEARS 3)

2022 ◽  
Author(s):  
Jiaqi Li ◽  
Zhaoyi He ◽  
Dongxue Li ◽  
Aichen Zheng

Abstract In order to improve the traffic safety of the tunnel pavement and reduce the impact of water seepage on the pavement structure, a convolutional neural network (CNN) model is established based on image detection technology to realize the identification, classification and statistics of pavement seepage. First, compared with the MobileNet network model, the deep learning model EfficientNet network model was built, and the accuracy of the two models was analyzed for pavement seepage recognition. The F1 Score was introduced to evaluate the accuracy and comprehensive performance of the two models for different types of seepage characteristics. Then the three gray processing methods, six threshold segmentation methods, as well as three filtering methods were compared to extract water seepage characteristics of digital image. Finally, based on the processed image, a calculation method of water seepage area was proposed to identify the actual asphalt pavement water seepage. The result shows that the recognition accuracy of the EfficientNet network model in the training set and the validation set are 99.85% and 97.53%, respectively, and the prediction accuracy is 98.00%. The accuracy of pavement water seepage recognition and prediction is better than the MobileNet network model. Using the cvtColor function for gray processing, using THRESH_BINARY for threshold segmentation, and using a combination of median filtering and morphological opening operations for image noise reduction can effectively extract water seepage characteristics. The water seepage area calculated by the proposed method has a small difference with the actual water seepage area, and the effect is agreeable.


2022 ◽  
Vol 2022 ◽  
pp. 1-10
Author(s):  
Shenming Yu

The study focused on the extraction of cardiovascular two-dimensional angiography sequences and the three-dimensional reconstruction based on the local threshold segmentation algorithm. Specifically, the two-dimensional cardiovascular angiography sequence was extracted first, and Gaussian smoothing was adopted for image preprocessing. Then, optimize maximum between-class variance (OSTU) was compared with the traditional two-dimensional OSTU and fast two-dimensional OSTU and applied in the segmentation of cardiovascular angiography images. It was found that the cardiovascular structure itself was continuous, the contrast agent diffused relatively evenly in the blood vessel, and the gray level of the blood vessel was also continuous. The degree of smoothness was consistent in all directions by Gaussian smoothing, avoiding the direction deviation of the smoothened image. The operation time (0.59 s) of the optimize OSTU was significantly shorter than that of traditional OSTU (35.68 s) and fast two-dimensional OSTU (6.34 s) ( P < 0.05 ). The local threshold segmentation algorithm can realize the continuous edge extraction of blood vessels and accurately reflect the stenosis of blood vessels. The results of blood vessel diameter measurement showed that the diameter from the end of blood vessel to the intersection varied linearly from 5.5 mm to 9.0 mm. In short, the optimize OSTU demonstrated good segmentation effects and fast calculation time; it successfully extracted continuous two-dimensional cardiovascular angiography images and can be used in three-dimensional reconstruction of cardiovascular images.


2021 ◽  
Author(s):  
Guang Yang ◽  
ZhaoGuang Liu ◽  
ZongNa Zhu

2021 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Swathi Kailasam ◽  
Sampath Dakshina Murthy Achanta ◽  
P. Rama Koteswara Rao ◽  
Ramesh Vatambeti ◽  
Saikumar Kayam

PurposeIn cultivation, early harvest offers farmers an opportunity to increase production while decreasing the chances of lower crop production rates, ensuring that the economy remains balanced. The significant reason is to predict the disease in plants and distinguish the type of syndrome with the help of segmentation and random forest optimization classification. In this investigation, the accurate prior phase of crop imagery has been collected from different datasets like cropscience, yesmodes and nelsonwisc . In the current study, the real-time earlier state of crop images has been gathered from numerous data sources similar to crop_science, yes_modes, nelson_wisc dataset.Design/methodology/approachIn this research work, random forest machine learning-based persuasive plants healthcare computing is provided. If proper ecological care is not applied to early harvesting, it can cause diseases in plants, decrease the cropping rate and less production. Until now different methods have been developed for crop analysis at an earlier stage, but it is necessary to implement methods to advanced techniques. So, the detection of plant diseases with the help of threshold segmentation and random forest classification has been involved in this investigation. This implemented design is verified on Python 3.7.8 software for simulation analysis.FindingsIn this work, different methods are developed for crops at an earlier stage, but more methods are needed to implement methods with prior stage crop harvesting. Because of this, a disease-finding system has been implemented. The methodologies like “Threshold segmentation” and RFO classifier lends 97.8% identification precision with 99.3% real optimistic rate, and 59.823 peak signal-to-noise (PSNR), 0.99894 structure similarity index (SSIM), 0.00812 machine squared error (MSE) values are attained.Originality/valueThe implemented machine learning design is outperformance methodology, and they are proving good application detection rate.


Plant Methods ◽  
2021 ◽  
Vol 17 (1) ◽  
Author(s):  
Xihuizi Liang

Abstract Background Cotton diceases seriously affect the yield and quality of cotton. The type of pest or disease suffered by cotton can be determined by the disease spots on the cotton leaves. This paper presents a few-shot learning framework that can be used for cotton leaf disease spot classification task. This can be used in preventing and controlling cotton diseases timely. First, disease spots on cotton leaf’s disease images are segmented by different methods, compared by using support vector machine (SVM) method and threshold segmentation, and discussed the suitable one. Then, with segmented disease spot images as input, a disease spot dataset is established, and the cotton leaf disease spots were classified using a classical convolutional neural network classifier, the structure and framework of convolutional neural network had been designed. At last, the features of two different images are extracted by a parallel two-way convolutional neural network with weight sharing. Then, the network uses a loss function to learn the metric space, in which similar leaf samples are close to each other and different leaf samples are far away from each other. In summary, this work can be regarded as a significang reference and the benchmark comparison for the follow-up studies of few-shot learning tasks in the agricultural field. Results To achieve the classification of cotton leaf spots by small sample learning, a metric-based learning method was developed to extract cotton leaf spot features and classify the sick leaves. The threshold segmentation and SVM were compared in the extracting of leaf spot. The results showed that both of these two method can extract the leaf spot in a good performance, SVM expented more time, but the leaf spot which extracted from SVM was much more suitable for classifying, thus SVM method can retain much more information of leaf spot, such as color, shape, textures, ect, which can help classficating the leaf spot. In the process of leaf spot classification, the two-way parallel convolutional neural network was established for building the leaf spot feature extractor, and feature classifier is constructed. After establishing the metric space, KNN was used as the spot classifier, and for the construction of convolutional neural networks, commonly used models were selected for comparison, and a spatial structure optimizer (SSO) is introduced for local optimization of the model, include Vgg, DesenNet, and ResNet. Experimentally, it is demonstrated that the classification accuracy of DenseNet is the highest, compared to the other two networks, and the classification accuracy of S-DenseNet is 7.7% higher then DenseNet on average for different number of steps. Conclusions As the step increasing, the accuracy of DesenNet, and ResNet are all improved, and after using SSO, each of these neural networks can achieved better performance. But The extent of increase varies, DesenNet with SSO had been improved the most obviously.


Sign in / Sign up

Export Citation Format

Share Document