Recently,the development and application of lane line departure warning systems have been in the *** any of the systems,the key part of lane line tracking,lane line identification,or lane line departure warning is whe...
详细信息
Recently,the development and application of lane line departure warning systems have been in the *** any of the systems,the key part of lane line tracking,lane line identification,or lane line departure warning is whether it can accurately and quickly detect lane *** 1990 s,they have been studied and implemented for the situations defined by the good viewing conditions and the clear lane markings on *** then,the accuracy for particular situations,the robustness for a wide range of scenarios,time efficiency and integration into higher-order tasks define visual lane line detection and tracking as a continuing research *** present,these kinds of lane marking line detection methods based on machine vision and imageprocessing can be divided into two categories:the traditional imageprocessing and semantic segmentation(includes deep learning)*** former mainly involves feature-based and model-based steps,and which can be classified into similarity-and discontinuity-based ones;and the model-based step includes different parametric straight line,curve or pattern *** semantic segmentation includes different machine learning,neural network and deep learning methods,which is the new trend for the research and application of lane line departure warning *** paper describes and analyzes the lane line departure warning systems,imageprocessingalgorithms and semantic segmentation methods for lane line detection.
Aiming at the limitations of traditional image recognition algorithms in the design of human resource management system, a design scheme of resource management system based on ant colony algorithm was proposed. Firstl...
详细信息
ISBN:
(数字)9798350382693
ISBN:
(纸本)9798350382709
Aiming at the limitations of traditional image recognition algorithms in the design of human resource management system, a design scheme of resource management system based on ant colony algorithm was proposed. Firstly, the influencing factors is accurately located through the colony foraging theory, and the indicators is reasonably divided to reduce interference, and the ant colony algorithm is used to construct the design scheme of the resource management system. Experimental results show that under certain evaluation criteria, the proposed scheme is superior to the traditional image recognition algorithm in terms of the design accuracy of the resource management system and the processing time of influencing factors, and has obvious advantages. The design of resource management system plays an extremely important role in human resources, which can accurately predict and optimize the growth characteristics and product generation of human resources. However, traditional image recognition algorithms have certain limitations in solving resource management simulation problems, especially when dealing with complex problems. In this paper, a resource management system design scheme based on ant colony algorithm is proposed to better solve this problem. In this scheme, the influencing factors were accurately located through the swarm foraging theory, so as to determine the division of indicators, and the ant colony algorithm was used to construct the scheme. Experimental results show that under certain evaluation criteria, the accuracy and speed of the scheme is significantly improved for different problems, and it has better performance. Therefore, the simulation scheme based on ant colony algorithm in the design of human resource management system can better solve the limitations of traditional image recognition algorithms and improve the simulation accuracy and efficiency.
Froth flotation is an important process in the mineral processing industry for extracting valuable materials. This work investigates online microscopic imaging and machine learning based image analysis methods for rea...
详细信息
ISBN:
(数字)9798331506230
ISBN:
(纸本)9798331506247
Froth flotation is an important process in the mineral processing industry for extracting valuable materials. This work investigates online microscopic imaging and machine learning based image analysis methods for real-time monitoring of the process. Previous limited work explored imaging the foam at the top surface layer of the froth flotation process. The new process imaging system in this work uses a corrosion-resistant online real-time imaging probe that can be put into the inside of the slurry and capture real-time images of bubbles. The acquired images are analyzed online using deep learning algorithms to automatically obtain key parameters of the bubbles, providing valuable data for froth flotation research and process control.
In this paper, we focus on the problem of bridge crack detection. In view of the characteristics of bridge cracks, such as the existence of multiple scales and shapes, and the interference of complex backgrounds and d...
详细信息
ISBN:
(数字)9798350389579
ISBN:
(纸本)9798350389586
In this paper, we focus on the problem of bridge crack detection. In view of the characteristics of bridge cracks, such as the existence of multiple scales and shapes, and the interference of complex backgrounds and different illumination conditions on the surface of bridges, we select the bridge crack images covering different scales, types, and illumination conditions to construct a dataset, and then carry out the image enhancement processes, such as rotating, randomly erasing, and adjusting the brightness and contrast of the cracks. Three deep learning algorithms, ResNet18, MobileNetV2 and EfficientNetB0, are selected for the experiments, and the performance of the models is evaluated by comparing the F1 scores, recall rates, confusion matrices, and observed loss curves. The results show that the proposed data processing and deep learning modeling strategy is effective in bridge crack detection, and the classification effect reaches a high level, providing a feasible method for bridge crack detection.
Sub-Nyquist sampling, also known as compressive sensing (CS), leverages the sparsity of ultrasound signals to reduce the sampling rate required for accurate ultrasound image reconstruction. Such under-sampling can red...
详细信息
ISBN:
(数字)9798350354959
ISBN:
(纸本)9798350354966
Sub-Nyquist sampling, also known as compressive sensing (CS), leverages the sparsity of ultrasound signals to reduce the sampling rate required for accurate ultrasound image reconstruction. Such under-sampling can reduce the computational complexity of the reconstruction algorithm while also relaxing on-board memory requirements. This paper demonstrates up to $10 \times$ reduction in total sample count by applying CS algorithms to ultrasound imaging using coherent plane-wave compounding (CPWC), and up to $57 \times$ reduction for single plane waves by combining CS with convolution beamforming. The resulting compression facilitates processing of the acquired samples directly on portable ultrasound devices, thus eliminating the need to transmit the data to a remote cloud for processing. The proposed compressed CPWC imaging approach is suitable for efficient target localization during focused ultrasound (FUS) neuromodulation. A portable ultrasound system for FUS neuromodulation, based on a system-on-chip (SoC) module and equipped with a dual-mode wearable probe, is also presented.
The operating environment of photovoltaic (PV) power plants is complex and continuously affected by dust soiling, particularly in arid and semi-arid regions of northwest China. Dust accumulation can reduce PV system e...
详细信息
ISBN:
(数字)9798331517724
ISBN:
(纸本)9798331517731
The operating environment of photovoltaic (PV) power plants is complex and continuously affected by dust soiling, particularly in arid and semi-arid regions of northwest China. Dust accumulation can reduce PV system efficiency, resulting in unstable energy output. To effectively detect and monitor dust impact on PV systems, this paper proposes a dust recognition method based on imageprocessing and machine learning. The method employs the Region Proposal Convolutional Neural Network (R-CNN) approach to extract features from onsite images and classify them using a classification layer. The results demonstrate the high accuracy of the R-CNN algorithm in dust recognition under various environmental conditions, aiding in optimizing the operation and maintenance of PV systems.
Firstly, this paper studies the overall structure, key technologies and core hardware design of the high-resolution satellite image data receiving and processing system. The system adopts a real-time parallel image da...
详细信息
Firstly, this paper studies the overall structure, key technologies and core hardware design of the high-resolution satellite image data receiving and processing system. The system adopts a real-time parallel image data receiving and processing algorithm based on RS error correction coding. At the same time, according to the characteristics of a single chip, combined with software design, the system has made corresponding research on RS decoding, noise reduction and signal processing software, and has carried out effective scheme and simulation verification design for different needs. The system can provide users with an intuitive and accurate user experience by using real-time point to point data transmission and automatic analysis of processing results for calculation. image signal has rich and bright color expression and clarity. Compared with similar data sources, the spatial resolution is greatly improved. In traditional imageprocessing, channel coding is the simplest and most convenient coding method, but it is more serious for low noise and inter symbol interference; Now, the satellite image data processing is basically to compress and store the collected data, decode and compress them, and then upload them. In this way, the decoder and storage device do not need to work at the same time, and the computing resources are greatly increased.
Recently, machine learning algorithms have been widely used in the fields of imageprocessing, network security and natural language processing, etc., profoundly affecting human life. However, machine learning algorit...
Recently, machine learning algorithms have been widely used in the fields of imageprocessing, network security and natural language processing, etc., profoundly affecting human life. However, machine learning algorithms have the characteristics of uncertain output, vulnerability to adversarial attacks, and unexplained decision-making processes, which seriously threaten the security of machine learning-based face recognition, Malware detection, and autonomous driving. Hence, it is imperative for the security practitioners to evaluate algorithm security to ensure that security needs are met. In this article, the authors propose a set of security assessment index systems and methods for machine learning algorithms for image classification scenarios: Refer to the security specification of machine learning algorithms and requirements to construct the security index system of image classification model. Furthermore, The Analytic Network Process(ANP) is applied to quantify the index weights and the Technique for Order Preference by Similarity to an Ideal Solution(TOPSIS) is applied to screen the optimal model, and finally the sensitivity analysis is applied to prove the stability of the proposed method. Experimental results show that this method has certain value and effect in assessing the security and model screening of image classification models.
This paper combines the advantages of Convolutional Neural Network (CNN), Long Short-Term Memory Network (LSTM), and A* search algorithm in order to achieve efficient path planning for intelligent vehicles or robots i...
详细信息
ISBN:
(数字)9798350360240
ISBN:
(纸本)9798350384161
This paper combines the advantages of Convolutional Neural Network (CNN), Long Short-Term Memory Network (LSTM), and A* search algorithm in order to achieve efficient path planning for intelligent vehicles or robots in complex environments. In this framework, CNN is responsible for extracting key spatial features from environment-aware data by learning and recognizing potential obstacles, drivable areas, and other spatial information. LSTM, on the other hand, is used to process time-series data and make predictions based on historical path information and real-time dynamic changes in order to take into account possible future changes in the environmental state. Subsequently, the environment model is generated on the basis of CNN and LSTM and fed into the A* search algorithm. The A* algorithm utilizes a heuristic strategy that combines the actual cost with the estimated cost in order to efficiently explore and find the optimal path. It aims to optimize multiple performance metrics such as path length, security and real-time performance, and provide adaptable and efficient path planning solutions for the intelligences while satisfying the constraints.
It is necessary to regularly detect faults to maintain the safety and stability of power lines. Insulators are one of the important electrical components in high-voltage transmission lines. It is extremely necessary t...
详细信息
ISBN:
(数字)9798350357882
ISBN:
(纸本)9798350357899
It is necessary to regularly detect faults to maintain the safety and stability of power lines. Insulators are one of the important electrical components in high-voltage transmission lines. It is extremely necessary to check the working status of insulators regularly. Traditional manual inspection is inefficient because it requires a significant amount of labor costs. In this paper, a method for detecting insulators' missing defect based on aerial images is proposed to address the issue by unmanned aerial vehicle (UAV). Firstly, the improved Faster R-CNN (region-based convolutional neural network) is used to identify and locate insulators in aerial images. Secondly, the U-Net image segmentation network segments insulators from the images. The adaptive threshold segmentation method completely separates the insulator from the background. Then the binary image of the insulator is obtained. Finally, the binary image is converted into a fault curve which is used for determining the missing insulators based on the distribution of the fault curve. By using collected insulator datasets on a 330kV overhead transmission line using a DJI M300 UAV platform and an onboard H20T camera/sensor, the detection accuracy of glass insulators is as high as 0.98 with the proposed algorithm. The positioning accuracy of the proposed algorithm is also higher than other algorithms. This method has high detection accuracy for missing defects in insulators. The experimental results show that compared with similar algorithms, this method has higher accuracy and efficiency.
暂无评论