Computers and Electronics in Agriculture最新文献

{"title":"A novel dual-branch spatial-spectral attention fusion model and method: A case study for the detection of nicotine content in tobacco leaves","authors":"Fukang Xing ,&nbsp;Rongguang Zhu ,&nbsp;Shichang Wang ,&nbsp;Lingfeng Meng ,&nbsp;Fujia Dong ,&nbsp;Songfeng Wang ,&nbsp;Jie Ren ,&nbsp;Zongxiu Bai ,&nbsp;Yapeng Kang","doi":"10.1016/j.compag.2025.110610","DOIUrl":"10.1016/j.compag.2025.110610","url":null,"abstract":"<div><div>Hyperspectral imaging (HSI) is a powerful tool for crop phenotypic component analysis, but developing efficient collaborative extraction and modeling methods for image and spectral features is a challenge. This study proposed a novel spatial-spectral fusion detection method for nicotine content utilizing hyperspectral imaging (HSI) and deep learning. Spectra of different regions and multi-channel images extracted by two-dimensional correlation analysis (2D-COS) were employed as inputs. A dual-branch spatial-spectral attention fusion model (DSSAM) was developed to enhance the expression ability of different modal information. Among them, two branches designed based on the residual module were used to extract spatial and spectral features, respectively. For the spectral branch, a multi-region spectral attention encoder (MSAE) was added to dynamically adjust the weights of the spectrum across leaf regions. For the spatial branch, a swin window attention (SWA) module was introduced to improve local feature extraction and spatial structure learning. The results demonstrated that MSAE and SWA could improve the spatial-spectral information fusion ability of the DSSAM. Compared with the dual-branch model without the attention modules, the coefficient of determination (R²) and relative prediction deviation (RPD) of the DSSAM model on the test set increased by 7.85% and 2.64%, respectively, and the Root Mean Square Error (RMSE) decreased by 39.29%. In addition, the DSSAM outperformed traditional chemometric and single-modal models, with a R² of 0.893, a RMSE of 0.289, and a RPD of 3.054. These findings provide a valuable approach for the quality nondestructive detection of cured tobacco leaves and other crop phenotypic components.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"236 ","pages":"Article 110610"},"PeriodicalIF":7.7,"publicationDate":"2025-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144185386","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"CBFW-YOLOv8: Automated recognition method for fish body surface diseases in recirculating aquaculture systems","authors":"Yihan Yin ,&nbsp;Xueqian Sun ,&nbsp;Guanghui Yu ,&nbsp;Jiayi Wang ,&nbsp;Daoliang Li ,&nbsp;Yang Wang","doi":"10.1016/j.compag.2025.110612","DOIUrl":"10.1016/j.compag.2025.110612","url":null,"abstract":"<div><div>Timely and accurate recognition of fish diseases is crucial for enhancing fish welfare and minimizing economic losses in aquaculture facilities. However, previous model-based detection mostly focused on detecting single fish diseases under ideal conditions, ignoring the impact of complex background factors and mutual occlusion of fish schools on model detection performance in actual aquaculture scenarios. Therefore, to automatically identify the health status of fish, we propose a high-precision, lightweight recognition method, CBFW-YOLOv8, specifically designed for detecting body surface diseases in fish within factory recirculating aquaculture systems. First, an underwater image acquisition platform is developed to collect a comprehensive dataset of spotted knifejaw diseases, including skin ulceration and tail rot disease. Subsequently, the Automatic Color Enhancement (ACE) algorithm is utilized to enhance the clarity and resolution of the original underwater images. In addition, the YOLOv8 backbone network, which features high computational complexity, is replaced by ConvNeXt V2 network. Futhermore, Focal Modulation and Bi-directional Feature Pyramid Network (BiFPN) feature fusion network are introduced to realize the fusion of low-level features and high-level semantic information, so that model can extract more abundant feature information. Finally, the Wise-IoU (weighted interpolation of sequential evidence for intersection over union) loss function is utilized to replace the CIoU (Complete over Union) loss function to improve the network boundary box regression performance and the detection effect of small target diseases. The experimental findings indicate that the CBFW-YOLOv8 model achieves a mean accuracy ([email protected]) of 93 % for detecting diseases in spotted knifejaw, with a detection speed of 50.2f/s and a model size of 13 M. When compared to alternative object detection models such as Faster-RCNN, SSD, YOLOv6, YOLOv8, and YOLOv10, the proposed approach demonstrates superior overall performance in detection precision and processing speed. The CBFW-YOLOv8 model had a [email protected] of 93 % and 97.6 %, a precision of 93.6 % and 96.2 %, and a recall of 89.3 % and 95.2 % for largemouth bass disease detection and floating head behavior detection, respectively. These metrics indicate that the model exhibits strong generalization capabilities. The developed method enables rapid, non-destructive detection of fish surface diseases, providing comprehensive technical assistance for the implementation and utilization of smart inspection robots in aquaculture facilities.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"236 ","pages":"Article 110612"},"PeriodicalIF":7.7,"publicationDate":"2025-05-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144185385","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"WLUSNet: A lightweight wheat lodging segmentation network based on UAV image","authors":"Qilei Zhu ,&nbsp;Ke Wang ,&nbsp;Dong Liang ,&nbsp;Jun Tang","doi":"10.1016/j.compag.2025.110587","DOIUrl":"10.1016/j.compag.2025.110587","url":null,"abstract":"<div><div>Wheat lodging is a usual agricultural disaster in wheat growth. It reduces the grain yield and harvesting efficiency. Existing segmentation methods cannot achieve satisfactory performance and trade-offs between accuracy, inference time, and lightweight when facing the challenge of multiple lodging scenes. Therefore, developing an innovative segmentation algorithm that is real-time and low-complexity to identify lodging situations is of great value for improving agricultural production. To achieve these goals, we propose a lightweight and efficient lodging semantic segmentation model, WLUSNet, to separate the lodging area of unmanned aerial vehicle (UAV) images. Inspired by the mixed depth-wise grouping convolution (MC) and the channel feature pyramid (CFP) modules, a multiscale backbone (MC-CFP) is designed to reduce information loss in feature extraction. Then, drawing on the characteristics of MC and the channel attention (CA) mechanism, a space pyramid module (MC-SP) is designed to enhance feature representation by obtaining the global information on the channel feature and the local information on the space feature. To reconstruct a high-resolution feature map, a feature fusion module (EDFF) between the shallow and deep features is introduced to improve segmentation accuracy. The comprehensive experimental results demonstrate that WLUSNet performs excellently well compared with 11 other state-of-the-art (SOTA) segmentation algorithms. WLUSNet achieves a mean intersection over union (mIoU) of 86.9, a mean pixel accuracy (mPA) of 93.26, a model size of 4.1 M, and an inference speed of 26.94 FPS on the self-built UAV remote sensing dataset in this paper. The generation experiment indicates that WLUSNet has the potential to segment other lodging crops, and can provide technical support for segmentation tasks in crop lodging.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"237 ","pages":"Article 110587"},"PeriodicalIF":7.7,"publicationDate":"2025-05-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144184446","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Numerical simulation and optimization design of a novel longitudinal-flow online fertilizer mixing device","authors":"Kang Zheng ,&nbsp;Shuo Yang ,&nbsp;Yuanyuan Gao ,&nbsp;Xiu Wang ,&nbsp;Jiakai Wang ,&nbsp;Senlin Song ,&nbsp;Changyuan Zhai ,&nbsp;Liping Chen","doi":"10.1016/j.compag.2025.110546","DOIUrl":"10.1016/j.compag.2025.110546","url":null,"abstract":"<div><div>To enhance the mixing uniformity and real-time performance of orchard fertilizer blending and application systems based on prescription fertilization maps, a novel longitudinal-flow online Fertilizer Mixing Device (FMD) was designed based on the convective mixing mechanism. Using the Discrete Element Method (DEM), an numerical model of structural parameters for the FMD was established. Moreover, a Central Composite Design (CCD) simulation experiment was conducted, with blending blade speed (A), blade pitch (B), and blade width (C) as experimental factors. The Coefficients of Variation (CV) for the urea (N), diammonium phosphate (P), and potassium sulfate (K) mixing fertilizers were selected as response indicators. A quadratic polynomial regression model was fitted to describe the relationship between experimental factors and indicators. Through model optimization, a group of optimal operating parameters for the device were determined to be: A = 1200 rpm, B = 90 mm, and C = 17 mm. The research introduced the Lacey index to evaluate mixing performance. Experimental results revealed that the fertilizer in the device reached a highly uniform mixing state within a response time of 1 s. Bench-scale validation tests under optimal parameters showed that the FMD achieved performance indicators of CV_N = 8.54 %, CV_P = 9.04 %, and CV_K = 10.61 %. The error between the experimental results and the simulation model was less than 5 %, indicating high predictive accuracy of the model. The findings provide valuable references for the mechanical structure design and parameter optimization of orchard online fertilizer blending and application systems based on prescription maps.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"237 ","pages":"Article 110546"},"PeriodicalIF":7.7,"publicationDate":"2025-05-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144169892","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A 3D group object tracking method for honeybees in open spaces","authors":"Jing Hu ,&nbsp;Yifan Chen ,&nbsp;Hongzhi Zhang ,&nbsp;Yiqiang Zhang ,&nbsp;Zican Shi ,&nbsp;Jie Ren ,&nbsp;Hengkang Ye ,&nbsp;Zhiyong Zuo ,&nbsp;Zhenbao Luo","doi":"10.1016/j.compag.2025.110535","DOIUrl":"10.1016/j.compag.2025.110535","url":null,"abstract":"<div><div>In this paper, we present a 3D group object stable tracking method based on binocular vision, called STBV, which is applied to the observation of flying honeybees in open spaces. Unlike typical multiple objects, group objects exhibit unique characteristics, such as a high population density, similar visual appearance among individuals, and similar motion patterns among individuals. Bees in free flight, as a representative example of such group objects, additionally possess traits such as small physical dimensions, agile motion, and variable target numbers, making effective tracking a formidable challenge. In this case, conventional trajectory construction strategies cause frequent ID switches, further causing instability of reconstructed 3D trajectories. To address this issue, the temporal stability of target instantaneous features and mutual support from binocular observation information are both used to improve the trajectory construction strategy in our paper. To verify our method, a 2D honeybee tracking dataset, HoneyBee2D, and a 3D honeybee tracking dataset, HoneyBee3D, were collected and annotated. The experimental results validate that the new strategy efficiently reduces the number of ID switches in 2D trajectories and consequently promotes the stability of reconstructed 3D trajectories. Furthermore, our reconstructed 3D flight trajectories of bees were used to analyze their motion and behavioral characteristics in a natural state and in an external disturbance state.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"237 ","pages":"Article 110535"},"PeriodicalIF":7.7,"publicationDate":"2025-05-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144169780","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MT-SRNet: A Transferable Multi-Task Super-Resolution Network for Pig Keypoint Detection, Segmentation, and Posture Estimation","authors":"Dong Liu,&nbsp;Andrea Parmiggiani,&nbsp;Tomas Norton","doi":"10.1016/j.compag.2025.110533","DOIUrl":"10.1016/j.compag.2025.110533","url":null,"abstract":"<div><div>Robust visual analysis of pigs in densely housed groups is essential for various applications within digital phenotyping and Precision Livestock Farming, including animal counting, posture recognition, tracking, body dimension measurement, and welfare assessment. However, the transferability and scalability of existing methods across diverse farming environments remain significant challenges. In this study, we addressed these issues by proposing a modular design that decouples environmental variations from individual animal variability. At the framework-level, we introduced a Rotated Bounding Box (RBB) detector combined with an Auto-Visual Prompt strategy to effectively suppress background interference and achieve seamless scene adaptation. At the task-level, we developed a Multi-task Super-Resolution Network (MT-SRnet) capable of simultaneously predicting pig keypoints, masks, and postures from low-resolution inputs. Our experimental results demonstrate that MT-SRnet maintains high accuracy (Keypoints – 95.33 % mPCK; Mask – 95.53 % mIoU; posture – 89.72 % accuracy) while substantially reducing model complexity by over 30-fold. Moreover, the proposed approach achieves a substantial increase in inference speed (up to ∼17,000 pigs/s), highlighting its practical suitability for real-time monitoring under resource-limited conditions. More information available online: <span><span>https://gitlab.kuleuven.be/m3-biores/public/m3pig</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"237 ","pages":"Article 110533"},"PeriodicalIF":7.7,"publicationDate":"2025-05-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144169890","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"3D multimodal image registration for plant phenotyping","authors":"Eric Stumpe ,&nbsp;Gernot Bodner ,&nbsp;Francesco Flagiello ,&nbsp;Matthias Zeppelzauer","doi":"10.1016/j.compag.2025.110538","DOIUrl":"10.1016/j.compag.2025.110538","url":null,"abstract":"<div><div>The use of multiple camera technologies in a combined multimodal monitoring system for plant phenotyping offers promising benefits. Compared to configurations that rely on a single camera technology, cross-modal patterns can be recorded that allow a more comprehensive assessment of plant phenotypes. However, the effective utilization of cross-modal patterns depends on image registration to achieve pixel-precise alignment - a challenge often complicated by parallax and occlusion effects inherent in plant canopy imaging. In this study, we propose a novel multimodal 3D image registration method that addresses these challenges by integrating depth information from a time-of-flight camera into the registration process. By leveraging depth data, our method mitigates parallax effects, facilitating more accurate pixel alignment across camera modalities. Additionally, we introduce an automated mechanism to identify and differentiate various types of occlusions, thereby minimizing registration errors. To evaluate the efficacy of our approach, we conduct experiments on a diverse dataset comprising six distinct plant species with varying leaf geometries. Our results demonstrate the robustness of the proposed registration algorithm, showcasing its ability to achieve accurate alignment across different plant types and camera compositions. Compared to previous methods our approach is not reliant on detecting plant-specific image features, making it suitable for a wide range of applications in plant sciences. Moreover, the registration approach can scale to arbitrary numbers of cameras with varying resolutions and wavelengths. Overall, our study contributes to advancing the field of plant phenotyping by offering a robust and reliable solution for multimodal image registration.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"237 ","pages":"Article 110538"},"PeriodicalIF":7.7,"publicationDate":"2025-05-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144169778","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multiscale phenotyping of grain crops based on three-dimensional models: A comprehensive review of trait detection","authors":"Jiangtao Qi ,&nbsp;Fangfang Gao ,&nbsp;Yang Wang ,&nbsp;Weirong Zhang ,&nbsp;Sisi Yang ,&nbsp;Kangkang Qi ,&nbsp;Ruirui Zhang","doi":"10.1016/j.compag.2025.110597","DOIUrl":"10.1016/j.compag.2025.110597","url":null,"abstract":"<div><div>Crop phenotyping is a reliable method for achieving crop breeding improvement, which can provide information for effective agricultural management and crop variety identification. Due to the significant differences in phenotypic characteristics of crops at the population, individual, and organ levels, it is still challenging to quickly and accurately complete large-scale crop phenotyping. Increasingly, phenotyping systems based on three-dimensional (3D) models have been continuously developed, making automated, high-precision data collection and fine-grained trait measurement possible. A review is performed to investigate and analyze the research work regarding multi-scale phenotyping of grain crops based on three-dimensional (3D) models. The focus is on traits that can evaluate the morphological changes of grain crops including maize, wheat, rice, soybean, and sorghum. Besides, the application of high-throughput phenotyping platforms, advanced sensor technologies, and artificial intelligence in phenotypic data processing is highlighted. New trends in the complex interaction of gene-environment-phenotype are revealed. Finally, research progress, major challenges and future prospects were discussed, hoping to provide a certain perspective for researchers and researchers preparing to enter this field.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"237 ","pages":"Article 110597"},"PeriodicalIF":7.7,"publicationDate":"2025-05-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144168250","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Development of nutrient management system based on ion-equivalent concentration ratio and ion monitoring to maintain ionic balance in closed hydroponic solution","authors":"Min-Seok Gang ,&nbsp;Hak-Jin Kim ,&nbsp;Tae In Ahn ,&nbsp;Woo-Jae Cho ,&nbsp;Sang-Hyun Lee ,&nbsp;Ju Young Lee ,&nbsp;Ji-Eun Hwang","doi":"10.1016/j.compag.2025.110588","DOIUrl":"10.1016/j.compag.2025.110588","url":null,"abstract":"<div><div>An automated nutrient management system was developed to maintain the ionic balance in the reused nutrient solution in a closed-loop hydroponics system. Fertilizers were supplied with variable compositions, using a replenishment algorithm based on electrical conductivity (EC) and ion-equivalent concentration ratio monitored with ion-selective electrodes (ISEs). An array of NO₃^–, K⁺, and Ca²⁺ electrodes was fabricated to measure ion concentrations in the drained solution. The measured ion concentrations were converted into ion-equivalent concentration ratios and fed back for adaptive control to determine the fertilizer composition. The system adjusted the fertilizer composition twice a week, using fertilizers (Ca(NO₃)₂·4H₂O, KH₂PO₄, KNO₃, NH₄NO₃, MgSO₄·7H₂O, K₂SO₄) and an acid solution (H₃PO₄) for pH adjustment. The fertilizer volumes were calculated using nonlinear programming methods based on the fertilizer compositions. The system activated each dosing channel differently to introduce fertilizers into the mixing tank. The system’s performance was tested by growing lettuce (<em>Lactuca sativa</em> L.) in a greenhouse. The measured ion concentrations were compared with the results of chemical analysis, and the actual ion-equivalent concentration ratios of the drained solution were compared to the target ratio of a standard Hoagland solution. The root mean square errors (RMSEs) of NO₃^–, K⁺, and Ca²⁺ measurements using ISEs were 39.18 mg L⁻¹, 17.79 mg L⁻¹, and 20.64 mg L⁻¹, respectively. The RMSE of the target ratios on the last day of cultivation was 5.9 %, representing a 58 % improvement compared to the system based on only EC. Therefore, the developed system may complement the conventional EC-based system for closed hydroponics.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"237 ","pages":"Article 110588"},"PeriodicalIF":7.7,"publicationDate":"2025-05-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144169891","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Real-Time agricultural image encryption algorithm using AES on edge computing devices","authors":"Mohammad Ashik Alahe ,&nbsp;Young Chang ,&nbsp;James Kemeshi ,&nbsp;Kwanghee Won ,&nbsp;Xufei Yang ,&nbsp;Lin Wei","doi":"10.1016/j.compag.2025.110594","DOIUrl":"10.1016/j.compag.2025.110594","url":null,"abstract":"<div><div>Smart agricultural industries increasingly rely on image-based data containing sensitive information about crops, such as nutrient levels, irrigation, disease, and more collected from the field. Protecting this critical data from unauthorized access from the field is essential to ensure the integrity of modern farm management systems. Image encryption is a crucial method to maintain data confidentiality during transmission from the field to secured data storage. While various image encryption techniques exist, limited attention has been given to securing real-time agricultural image data. Our work proposes a novel, cost-effective, real-time agricultural image encryption technique to protect sensitive agricultural data from potential cyberattacks during transmission. Using the Advanced Encryption Standard algorithm on edge computing devices like Jetson Nano and Jetson Xavier NX, agricultural images were encrypted and immediately transferred from the field to a base computer for further processing. Results show high sensitivity to any bit change, high MSE, and low PSNR values. The NPCR and UACI values are greater than the minimum value of 99.6094% and 33.4635%, respectively, confirming its robustness against statistical and differential attacks. Additionally, the performance of both devices was assessed across several key metrics, to determine their suitability for securing real-time agricultural images. The experimental results demonstrated that the Jetson Xavier NX consistently outperformed the Jetson Nano. This difference was particularly noticeable at higher frame rates, where the Jetson Xavier NX exhibited a substantial improvement in both decryption and total processing times compared to the Jetson Nano.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"237 ","pages":"Article 110594"},"PeriodicalIF":7.7,"publicationDate":"2025-05-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144169779","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}