Remote sensing object detection faces significant challenges due to the varying resolutions of images and diverse shapes objects. In this paper, we introduce an innovative approach enhance networks in remote by addressing these issues through a novel two-stage training process. Our key contributions include development combined low-resolution pretraining high-resolution fine-tuning strategy, which significantly improves model's generalization capability across different resolutions.We have...

<title>Abstract</title> As Connected and Automated Vehicle (CAV) technology progresses, optimizing vehicle-road cooperative control becomes critical for alleviating traffic congestion. Existing approaches often involve complex computations that are unsuitable real-time typically limited to isolated intersections, restricting their scalability. To address these limitations, this paper introduces a Mixed Platoon Dual-Layer (MPDL) model designed large-scale road networks. The leverages mixed...

On the basis of symmetric encoding algorithm for modified signed-digit (MSD), a 7*7 truth table that can be realized with optical methods was developed. And based on table, path structures and circuit implementations one-step MSD adder ternary computer (TOC) were designed. Experiments show scheme is correct, feasible, efficient.

Leaf spot (LS) caused by Cercosporidium personatum is one of the most harmful peanut diseases in late growth stage and severely affects yield peanuts. Hyperspectral disease detection technology efficient, objective, accurate suitable for large-scale crop management practices. To establish a multi-scale spectral index (SI) with high accuracy stability LS disease, reflectance different severity levels at leaf, plant, field scales was collected, difference wavelength analyzed using mean,...

Abstract Thick-walled steel pipes, which bear high internal pressure, are widely applied in nuclear power and pressure pipelines. If there defects the inner wall, they easy to expand cause accidents. Therefore, thick-walled pipe must be subject non-destructive testing after production. For magnetic flux leakage (MFL) method, detection sensitivity gradually decreases with increase of wall thickness. To solve this problem, a new structure MFL probe is proposed paper. The influence iron core...

Hi-lock bolt fasteners are widely used in aircraft multilayer structures, and it is especially important to detect hole edge defects of hi-lock connection structures. In this study, an array remote field eddy current (EC) testing system based on a field-programmable gate (FPGA) EC probe multiplexing the principle were designed developed. Using system, experiments carried out fastener test blocks, effects defect length buried depth detection signal studied. The experimental results show that...

Remote sensing image classification plays a crucial role in the field of remote interpretation. With exponential growth multi-source data, accurately extracting target features and comprehending attributes from complex images significantly impacts accuracy. To address these challenges, we propose Canny edge-enhanced multi-level attention feature fusion network (CAF) for classification. The original is specifically inputted into convolutional extraction global features, while increasing depth...

Leaf spot disease is a dangerous that affects peanut growth, and its severity can significantly impact yield. Hyperspectral-based detection technology popular non-destructive technique due to high efficiency, objectivity, accuracy. In this study, leaf spectra at different levels of were collected in Liaoning Province, China, mid-August. This study analyzed the differences wavelengths using mean spectral reflectance sensitivity. Using improved principal component analysis loading (I-PCA...

In view of the common pests and diseases irregularly shaped fruits Nanguo pear, this paper fused spectral information image features to realize rapid nondestructive testing recognition external defects on pear by hyperspectral imaging technology. Backpropagation neural network support vector machine model was established identify defects, which are commonly used in classification pattern recognition. The results show that effect is better than backpropagation network. Among them, accuracy...

The RSA algorithm remains one of the most influential and extensively used public-key encryption algorithms in today’s cryptographic domain. However, traditional has several drawbacks, including slow computational speed, high cost, substantial space requirements practical applications. Consequently, this study amalgamates characteristics a ternary optical computer (TOC), parallelism, multi-digit precision, low energy consumption, to research propose dual-channel transmission strategy...

Embedded application is a hotspot at present and Linux gradually becomes the most important operating system for embedded applications. Aiming real-time problems of from five performance parameters system, this paper analyzes concludes that scheduling latency interrupt are fundamental constraints improving 2.6 kernel, then designs implements new task model operations to solve above problem. Hard algorithm which named as priority bitmap algorithm, response emphasized main codes given out....

With the reconfiguration technique, users can configure optical processor of a new computing platform, ternary computer (TOC), into 42 specific basic operation units, and reconfigure it when computation is completed. A simulative method software for this technique proposed in paper. Also, an elaborate experiment based on discussed. This simulation demonstrates that theory reconfiguring TOC valid also lays foundation hardware implementation technique.