Detection of Invasive Species (Siam Weed) Using Drone-Based Imaging and YOLO Deep Learning Model

Remote Sensing, Journal Year: 2025, Volume and Issue: 17(1), P. 120 - 120

Published: Jan. 2, 2025

This study explores the efficacy of drone-acquired RGB images and YOLO model in detecting invasive species Siam weed (Chromolaena odorata) natural environments. is a perennial scrambling shrub from tropical sub-tropical America that outside its native range, causing substantial environmental economic impacts across Asia, Africa, Oceania. First detected Australia northern Queensland 1994 later Northern Territory 2019, there an urgent need to determine extent incursion vast, rugged areas both jurisdictions for distribution mapping at catchment scale. tests drone-based imaging train deep learning contributes goal surveying non-native vegetation We specifically examined effects input training images, solar illumination, complexity on model’s detection performance investigated sources false positives. Drone-based were acquired four sites Townsville region test (YOLOv5). Validation was performed through expert visual interpretation results image tiles. The YOLOv5 demonstrated over 0.85 F1-Score, which improved 0.95 with exposure images. A reliable found be sufficiently trained approximately 1000 tiles, additional offering marginal improvement. Increased did not notably enhance performance, indicating smaller adequate. False positives often originated foliage bark under high low reduced these errors considerably. demonstrates feasibility using models detect landscapes, providing safe alternative current method involving human spotters helicopters. Future research will focus developing tools merge duplicates, gather georeference data, report detections large datasets more efficiently, valuable insights practical applications management

Language: Английский

---

Improving U-net network for semantic segmentation of corns and weeds during corn seedling stage in field

Frontiers in Plant Science, Journal Year: 2024, Volume and Issue: 15

Published: Feb. 9, 2024

Introduction Weeds are one of the main factors affecting crop growth, making weed control a pressing global problem. In recent years, interest in intelligent mechanical weed-control equipment has been growing. Methods We propose semantic segmentation network, RDS_Unet, based on corn seedling fields built upon an improved U-net network. This network accurately recognizes weeds even under complex environmental conditions, facilitating use weeding for reducing density. Our research utilized field-grown maize seedlings and accompanying expansive fields. integrated employing ResNeXt-50 feature extraction encoder stage. decoder phase, Layer 1 uses deformable convolution with adaptive offsets, replacing traditional convolution. Furthermore, concurrent spatial channel squeeze excitation is incorporated after ordinary convolutional layers Layers 2, 3, 4. Results Compared existing classical models such as U-net, Pspnet, DeeplabV3, our model demonstrated superior performance specially constructed grass dataset, CGSSD, during The Q6mean intersection over union (MIoU), precision, recall this 82.36%, 91.36%, 89.45%, respectively. to those original proposed achieves improvements 5.91, 3.50, 5.49 percentage points MIoU, recall, detection speed 12.6 frames per second. addition, ablation experiments further confirmed impactful contribution each improvement component overall performance. Discussion study provides theoretical technical support automated operation devices.

Language: Английский

---

Applications and perspectives of Generative Artificial Intelligence in agriculture

Computers and Electronics in Agriculture, Journal Year: 2025, Volume and Issue: 230, P. 109919 - 109919

Published: Jan. 10, 2025

Language: Английский

---

Medicinal and poisonous plants classification from visual characteristics of leaves using computer vision and deep neural networks

Ecological Informatics, Journal Year: 2024, Volume and Issue: 82, P. 102683 - 102683

Published: June 12, 2024

Poisonous plants are the third largest category of poisons known globally, which pose a risk poisoning and death to humans. Currently, identification medicinal poisonous is done by humans using experimental methods, not accurate associated with many errors, also use laboratory methods requires experts this method very costly time-consuming. Therefore, distinguishing between important emerging, non-destructive, fast such as computer vision artificial intelligence. In study, we propose robust generalized model spatial attention (SA) channel (CA) modules for classification different plants. A dataset containing 900 confirmed images three plant classes (oregano, weed) was used. The mechanisms enhance efficiency deep learning (DL) networks allowing them precisely focus on all relevant input elements. order performance proposed model, CA implemented based four pooling operations including global average pooling-based (GAP-CA), mixed (Mixed-CA), gated (Gated-CA), tree (Tree-CA) operations. results showed that DL Tree-CA had promising outperformed other state-of-the-art models, achieving values 99.63%, 99.38%, 99.52%, 99.74%, 99.42%, accuracy, precision, recall, specificity, F1-score, respectively. findings support our model's success in identifying from similar Recent advancements computer-based technologies intelligence enable automatic detection plants, revolutionizing traditional methods.

Language: Английский

---

Object-level benchmark for deep learning-based detection and classification of weed species

Crop Protection, Journal Year: 2023, Volume and Issue: 177, P. 106561 - 106561

Published: Dec. 17, 2023

Weeds can decrease yields and the quality of crops. Detection, localisation, classification weeds in crops are crucial for developing efficient weed control management systems. Deep learning (DL) based object detection techniques have been applied various applications. However, such generally need appropriate datasets. Most available datasets only offer image-level annotation, i.e., each image is labelled with one species. practice, multiple (and crop) species and/or instances Consequently, lack instance-level annotations puts a constraint on applicability powerful DL techniques. In current research, we construct an dataset. The images sourced from publicly dataset, namely Corn It has 5997 plants four types weeds. We annotated dataset using bounding box around instance them crop or weed. Overall, contain about three average, while some over fifty boxes. To establish benchmark evaluated several models, including YOLOv7, YOLOv8 Faster-RCNN, to locate classify performance models was compared inference time accuracy. YOLOv7 its variant YOLOv7-tiny both achieved highest mean average precision (mAP) 88.50% 88.29% took 2.7 1.43 ms, respectively, image. YOLOv8m, YOLOv8, detected 2.2 ms mAP 87.75%. Data augmentation address class imbalance improves results 89.93% 89.39% YOLOv8. accuracy performed by this research indicate that these be used develop automatic field-level system.

Language: Английский

---

Advancing Precision Agriculture: Enhanced Weed Detection Using the Optimized YOLOv8T Model

Arabian Journal for Science and Engineering, Journal Year: 2024, Volume and Issue: unknown

Published: Aug. 2, 2024

Language: Английский

---

Multiscale feature fusion and enhancement in a transformer for the fine-grained visual classification of tree species

Ecological Informatics, Journal Year: 2025, Volume and Issue: unknown, P. 103029 - 103029

Published: Jan. 1, 2025

Language: Английский

---

A REVIEW OF INNOVATIVE DESIGN AND INTELLIGENT TECHNOLOGY APPLICATIONS OF THRESHING DEVICES IN COMBINE HARVESTERS FOR STAPLE CROPS

INMATEH Agricultural Engineering, Journal Year: 2025, Volume and Issue: unknown, P. 706 - 725

Published: April 28, 2025

This paper reviews the progress in innovative design and intelligent technology applications of threshing devices combine harvesters for staple crops. To address issues poor adaptability low intelligence traditional systems, researchers have significantly improved performance by optimizing components drum structures. Meanwhile, machine vision deep learning achieved important breakthroughs feed rate monitoring, breakage impurity detection, control. review aims to provide a reference research system structural optimization operational parameter

Language: Английский

---

A Review on Crop and Weed Identification Using Deep Learning

INTERANTIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT, Journal Year: 2024, Volume and Issue: 08(04), P. 1 - 5

Published: April 26, 2024

Agriculture remains the backbone of several economies in world, especially underdeveloped countries. With rapid growth population and increasing demand for food, farmers need to maximize productivity one possibility is reduction losses. Weeds are major dangers farming. Indeed, they compete vigorously with crop nutrients water. Improved methods required get good yields from crops. The proposed model aims organize a diverse dataset weed images, leveraging Convolutional Neural Networks (CNN) identification, texture feature extraction, employing CNN precise identification. paper introduces novel deep-learning approach utilizing Residual (ResNet) effectively identifying classifying crops weeds agriculture. findings underscore effectiveness various deep learning models such as CNNs, accurately detecting within crops, aided by preprocessing techniques optimization. These advancements hold promising prospects revolutionizing agricultural practices enhancing future. Key Words: Agriculture, Weed Management, Crop Identification, Networks, Deep Learning

Language: Английский

---

Image Analysis Artificial Intelligence Technologies for Plant Phenotyping: Current State of the Art

AgriEngineering, Journal Year: 2024, Volume and Issue: 6(3), P. 3375 - 3407

Published: Sept. 17, 2024

Modern agriculture is characterized by the use of smart technology and precision to monitor crops in real time. The technologies enhance total yields identifying requirements based on environmental conditions. Plant phenotyping used solving problems basic science allows scientists characterize select best genotypes for breeding, hence eliminating manual laborious methods. Additionally, plant useful such as subtle differences or complex quantitative trait locus (QTL) mapping which are impossible solve using conventional This review article examines latest developments image analysis AI, 2D, 3D reconstruction techniques limiting literature from 2020. collects data 84 current studies showcases novel applications various technologies. AI algorithms showcased predicting issues expected during growth cycles lettuce plants, soybeans different climates conditions, high-yielding improve yields. high throughput also facilitates monitoring crop canopies genotypes, root phenotyping, late-time harvesting weeds. methods combined with guide applications, leading higher accuracy than cases that consider either method. Finally, a combination undertake operations involving automated robotic harvesting. Future research directions where uptake smartphone-based time series ML recommended.

Language: Английский

---