Bolt defect inspection is an important work in transmission line inspection. Due to the small size of bolts images, existing algorithms are difficult extract valuable features and achieve poor performance on bolt detection. This paper proposed ultra-small detection model(UBDDM) based a deep convolutional neural network(DCNN), including object perception module(UOPM) local module(LBDM). In this paper, UOPM first constructed realize coarse region recognition for salient high-resolution image...

Plant disease recognition is of vital importance to monitor plant development and predicting crop production. However, due data degradation caused by different conditions image acquisition, e.g., laboratory vs. field environment, machine learning-based models generated within a specific dataset (source domain) tend lose their validity when generalized novel (target domain). To this end, domain adaptation methods can be leveraged for the learning invariant representations across domains. In...

By the means of a differential inequality technique, we obtain lower bound for blow‐up time if p and initial value satisfy some conditions. Also, establish criterion an upper under conditions as well nonblow‐up exponential decay other Copyright © 2014 John Wiley & Sons, Ltd.

Person re-identification (re-ID) performance has been significantly boosted in recent works, but re-ID model trained on one dataset usually cannot work effectively another one. To address this problem, we proposed a novel framework-double-domain translation generative adversarial network (DTGAN) that can train images between two domains and generalize the domain well to domain. We divide paper into steps. First, source are translated an unsupervised manner, retain style of target ID labels...

There is an irreversibility in the decline of Li-ion batteries, and performance individual cells battery pack will gradually as number times on-board charged discharged increases This situation can significantly affect daily use electric vehicles, for example by shortening driving range, addition, deterioration probability vehicle breakdowns. Very little work has been done on prediction lithium degradation long-mileage states, accurate future reduce EV failure, making very important. In this...

Icing monitoring system within internet of things is developed to provide multimodal data for assessing the degree icing on transmission lines. Nevertheless, existing methods relying solely either sensor or images demonstrate limited precision and inferior fault tolerance. This paper proposes a correlation-based feature fusion approach integrate both imaging data, thereby enhancing severity lines addressing aforementioned problems. The inherent correlation characteristics present in images,...

A novel CO2-capturing natural gas combined cycle is proposed, in which cold energy during LNG gasification can be fully utilized to reduce the power consumption of oxygen production and CO2 recovery process penalty. In a simulated example, with GE LMS100™ turbine corresponding 159 t·h−1 gasified LNG, electrical exergy efficiency proposed reaches 54.9%. same time, 90.6% could recovered. Compared without capture, has slight net loss, quantity avoided about 322 g·kW−1h−1. For plant cycle,...

The safe approach distance detection of large construction equipment in substations is important to ensure the safety and stability power system, as well prevent damage, outages other accidents. current method unable intelligently distinguish from realize real-time detection. Therefore, this paper constructs a system for large-scale based on stereo vision target detection, realizes high-precision between electric equipment. Firstly, distinguishes using GhostNet-based substation model....

Due to low radiation dose and software fault, CT images are noisy. How extraction of meaningful information from noisy challenging works. In this work, we present a new denoising algorithm for image using double density dual tree complex wavelet transform (DDCWT). Firstly, use DDCWT decompose into high frequency components. the next, modified threshold is used coefficient. Finally, denoised obtained by reconstructing components through inverse decomposition DDCWT. Experimental results...

Polluted insulators seriously threaten the safe and stable operation of power grids, which attaches great significance to insulator contamination perception. Among present methods, non-contact approaches based on infrared images have gradually been widely used, as they are much more low cost. However, thermal effect is largely affected by meteorological conditions, makes image-based methods less accurate. To solve above problem, we take image parameters including humidity temperature input,...

The power operation site typically faces high voltages, strong currents, and other environmental hazards. Safety fence as a boundary for is of great importance to ensure the safety operators. preparation temporary random, risk identification complex problem that involves dynamic interaction between individuals fence. existing methods determining area based on prior knowledge are insufficient. Therefore, this paper proposes method neighborhood correlation networks vector calculation. In...

Oxy-fuel combustion is a leading potential technology to capture CO 2 . Because the oxygen production process causing largest power penalty in oxy-fuel system, it essential cut down separation for capturing at low cost. This paper presents energy consumption analysis results of air units with three different cycles offering systems. The study shows that when gaseous compression (GOXC) cycle selected pressuring product stream, specific high pressure and purity triple column about 6.4-7.2%...

Robustness and security are difficult to be solved by conventional two-dimensional (2D) digital watermarking technology. In recent years, three-dimensional (3D) has become a new hotspot in optical information security. This paper presents improved 3D method based on computational integrated imaging cryptosystem. Firstly, is generated encrypted integral cryptosystem that implemented with smart pseudoscopic-to-orthoscopic conversion (SPOC) model. Secondly, discrete wavelet transform algorithm...

A novel watermarking approach for copyright protection of color images based on the nonsubsampled contour let transform (NSCT). We employ genetic algorithm optimization principles and HVS masking technique to obtain high robustness transparency with respect existing algorithms. In our scheme, embedding factor is controlled locally according result Generic Algorithm (GA) cycle. The simulation results show that proposed yields a watermark which invisible human eyes robust wide variety common attacks.

In this paper, a novel and effective steganalysis based on One-Class Support Vector Machine(OC-SVM) with Simulated Annealing clustering algorithm is proposed to blindly (i.e., without knowledge of the steganographic schemes) determine existence hidden messages in an image. The performance sample concerned OC-SVM multi-sphere. previous work, K-means mainly used create such multi-sphere by clustering. But traditional depends initial centers ends local minimum value. So, solve problem caused...