Research on Segmentation Method of Maize Seedling Plant Instances Based on UAV Multispectral Remote Sensing Images

Plants, Journal Year: 2024, Volume and Issue: 13(13), P. 1842 - 1842

Published: July 4, 2024

The accurate instance segmentation of individual crop plants is crucial for achieving a high-throughput phenotypic analysis seedlings and smart field management in agriculture. Current monitoring techniques employing remote sensing predominantly focus on population analysis, thereby lacking precise estimations plants. This study concentrates maize, critical staple crop, leverages multispectral data sourced from unmanned aerial vehicles (UAVs). A large-scale SAM image model employed to efficiently annotate maize plant instances, constructing dataset seedling segmentation. evaluates the experimental accuracy six algorithms: Mask R-CNN, Cascade PointRend, YOLOv5, Scoring YOLOv8, various combinations bands comparative analysis. findings indicate that YOLOv8 exhibits exceptional accuracy, notably NRG band, with bbox_mAP50 segm_mAP50 accuracies reaching 95.2% 94%, respectively, surpassing other models. Furthermore, demonstrates robust performance generalization experiments, indicating its adaptability across diverse environments conditions. Additionally, this simulates analyzes impact different resolutions model’s accuracy. reveal sustains high even at reduced (1.333 cm/px), meeting criteria.

Language: Английский

---

A Method for Quantifying Mung Bean Field Planting Layouts Using UAV Images and an Improved YOLOv8-obb Model

Agronomy, Journal Year: 2025, Volume and Issue: 15(1), P. 151 - 151

Published: Jan. 9, 2025

Quantifying planting layouts during the seedling stage of mung beans (Vigna radiata L.) is crucial for assessing cultivation conditions and providing support precise management. Traditional information extraction methods are often hindered by engineering workloads, time consumption, labor costs. Applying deep-learning technologies reduces these burdens yields reliable results, enabling a visual analysis distribution. In this work, an unmanned aerial vehicle (UAV) was employed to capture visible light images bean seedlings in field across three height gradients 2 m, 5 7 m following series approach. To improve detection accuracy, small target layer (p2) integrated into YOLOv8-obb model, facilitating identification seedlings. Image performance were analyzed considering various dates, heights, resolutions, K-means algorithm utilized cluster feature points extract row information. Linear fitting performed via least squares method calculate layout parameters. The results indicated that on 13th day post seeding, 2640 × 1978 image captured at above ground level exhibited optimal performance. Compared with YOLOv8, YOLOv8-obb, YOLOv9, YOLOv10, YOLOv8-obb-p2 model improved precision 1.6%, 0.1%, 0.3%, 2%, respectively, F1 scores 2.8%, 0.5%, 3%, respectively. This extracts information, data quantifying These findings can be rapid large-scale assessments growth development, theoretical technical counting hole-seeded crops.

Language: Английский

---

A statistical method for high-throughput emergence rate calculation for soybean breeding plots based on field phenotypic characteristics

Plant Methods, Journal Year: 2025, Volume and Issue: 21(1)

Published: March 24, 2025

In the process of smart breeding, rapid statistics soybean emergence rate, as an important part breeding screening, face challenges under environmental constraints, especially selection and varieties in dense environments. Due to influence factors, existing methods have shortcomings, such low throughput, efficiency, insufficient precision. Therefore, effective precise statistical method is required. this study, UAV (Unmanned Aerial Vehicle)-scale data combined with ground measurement were used research object explore feasibility improving accuracy screening intensive planting. To end, a set technical solutions, including background removal, detection, accurate counting, designed. Firstly, segmentation based on contrast enhancement filtering ultra-green eigenvalues Otsu algorithm was proposed remove complex remote sensing images retain morphological information seedlings. Secondly, deep learning detection model infer predict processed label Then, seedling counting constructed: by establishing growth model, idea "growth normalization" proposed, expansion-compression factor defined eliminate inconsistency counting. After in-depth analysis planting characteristics seedlings overlapping conditions, "inter-seedling occlusion algorithm" solve problem between order bounding box, soft strategy specially designed avoid redundant values brought it. Finally, according calculation results, thematic map rate plot plots displayed. experiments, can effectively count number image, overall 99.18% error 0.82%. addition, Yolov8n had best recognition effect task, mAP (0.5–0.95) 85.15%. The increased results 4.06%. It has been demonstrated through experimental tests verifications that solid support for work concerning condition provided method. This innovative played facilitating role accelerating also some new ideas reference directions further exploration efficient screening.

Language: Английский

---

A pumpkin seed vitality detection model based on deep spectral features

Computers and Electronics in Agriculture, Journal Year: 2025, Volume and Issue: 236, P. 110457 - 110457

Published: April 28, 2025

Language: Английский

---

3D terrestrial LiDAR for obtaining phenotypic information of cigar tobacco plants

Computers and Electronics in Agriculture, Journal Year: 2024, Volume and Issue: 226, P. 109424 - 109424

Published: Sept. 7, 2024

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

---