Leveraging smartphone photos and AI, a new computer vision framework accurately detects sorghum panicles and estimates grain counts with 17% error to accelerate field yield prediction. #ComputerVision #Phenomics
Plant Phenomics
@plantphenomics.bsky.social
An open access journal indexed in: #DOAJ #EI #PMC #SCIE (#JIF 2023: 7.6) #Scopus (#CiteScore2023: 8.6) etc. #PlantPhenomics #PlantPhenotyping #openaccess
Evaluating NeRFs for 3D plant reconstruction shows a 74.6% F1 score in outdoor fields. Early stopping halves training time with minimal accuracy loss. #Phenomics #NeRF Details: doi.org/10.34133/pla...
Virtual maize canopy simulations reveal that matching leaf area and nitrogen vertical distributions maximizes radiation use efficiency—laying groundwork for high-throughput phenotyping of elite cultivars. Details: doi.org/10.34133/pla...
AISOA-SSformer achieves 83.1% MIoU on rice leaf disease segmentation using sparse global-update perceptron, salient feature attention, and annealing-integrated sparrow optimization. Code: github.com/ZhouGuoXiong/Rice-Leaf-Disease-Segmentation-Dataset-Code 🌾 Details: doi.org/10.34133/pla...
Phenomic selection using parental NIR spectral data predicts hybrid rapeseed traits competitively with genomic selection—offering a zero-cost, high-throughput breeding tool already embedded in routine workflows. Details: doi.org/10.34133/pla...
New multimodal vision system predicts individual wheat anthesis 7–14 days ahead using RGB + weather data with few-shot learning, achieving >0.8 F1—revolutionizing breeding logistics and regulatory compliance. Details: doi.org/10.1016/j.pl...
RsegNet uses cosine feature extraction and dual-channel clustering to segment UAV LiDAR rubber tree point clouds, achieving 86.1% F-score—enabling precise structural monitoring for plantation management. #trees Details: doi.org/10.1016/j.pl...
New few-shot learning framework FSEA detects novel weed species with just 30 samples, outperforming SOTA detectors by embedding plant-specific prior knowledge (color, morphology, occlusion) into Faster R-CNN. 🌱🤖 Details: doi.org/10.1016/j.pl...
Low-cost (<€1,000) phenotyping reveals diverse quantitative disease resistance mechanisms in wild tomatoes, showing how temporal dynamics dissect complex host-pathogen interactions. #PlantPathology Details: doi.org/10.34133/pla...
LGNet: A dual-branch CNN+Transformer network with adaptive feature fusion achieves 88.74% on AI Challenger and 99.08% on corn disease datasets. #PlantDisease #DeepLearning #ComputerVision Details: doi.org/10.34133/pla...
This review surveys predictive modeling for plant growth forecasting, bridging deterministic, probabilistic, and generative approaches with dynamic environmental interactions for next-gen phenomics. 🌱📊 #PlantPhenomics #AI #Review Details: doi.org/10.1016/j.pl...
ART extracts 27 algorithmic root traits from images, outperforming traditional methods 96.3% vs 85.6% in wheat drought classification. 4 ARTs = 23 traditional traits. 🌾🤖 #PlantScience #Phenotyping #MachineLearning Details: doi.org/10.1016/j.pl...
CitrusGAN reconstructs 3D citrus CT models from just 6 sparse X-ray views, enabling high-throughput, low-cost fruit phenotyping with 92.1% structural similarity. 🍊🔬 #AI #Agriculture #Phenotyping Details: doi.org/10.1016/j.pl...
3D panoptic segmentation of apple & pear micro-CT tissue achieves AJI 0.89/0.77, outperforming 2D and watershed methods. First complete automated protocol for plant tissue morphometrics. 🍎🍐 #DeepLearning #PlantScience #CTImaging Details: doi.org/10.1016/j.pl...
A decade of deep learning in plant phenotyping: from first papers to a thriving research community bridging computer vision and biology. Our reflection on this transformative journey. 🌱🤖 #PlantPhenotyping #DeepLearning #AgTech Details: doi.org/10.1016/j.pl...
New GWFSS dataset enables pixel-level wheat organ segmentation (leaves, stems, spikes) across all growth stages. Segformer hits ~90% mIOU on leaves/spikes, but stems lag at 54%. 🌾 #AgTech #ComputerVision #AI Details: doi.org/10.1016/j.pl...
CitrusGAN reconstructs 3D citrus CT from just 6 sparse X-ray views with 92% similarity, enabling high-throughput, low-cost fruit phenotyping for breeding. 🍊🔬 Details: doi.org/10.1016/j.pl...
3D multispectral phenotyping reveals Chinese cabbage's rapid heat sensitivity & 4 distinct tolerance strategies, accelerating climate-resilient breeding via non-destructive trait screening. #HeatTolerance #CropBreeding Details: doi.org/10.1016/j.pl...
New multi-scale temporal deep learning framework fuses UAV imagery to track rice phenology across 500+ cultivars, cutting flight time 78% while maintaining 87.3% accuracy. #RiceBreeding #PrecisionAg Details: doi.org/10.1016/j.pl...
New SRD-YOLO detects weed growth points in corn fields with 96.5% mAP & 169 FPS, using lightweight AI to enable real-time precision weeding despite occlusion & variable lighting. #PrecisionAgriculture #WeedControl Details: doi.org/10.1016/j.pl...
New neural network generates synthetic 3D leaf point clouds from skeletons, boosting real-world trait estimation accuracy and reducing costly manual phenotyping labor. #PlantPhenotyping #AI Details: doi.org/10.1016/j.pl...
New study: Thinning boosts conifer light penetration (15%→22%) & carbon absorption (+18%) under drought. 3D canopy reconstruction reveals how forest structure drives climate resilience. #ForestEcology #ClimateChange Details: doi.org/10.1016/j.pl...
3D canopy photosynthesis model reveals light & physiology drive tomato yield gaps across climates. Key insight: leaf quantum efficiency beats plant spacing for boosting photosynthesis. 🍅 #PlantScience #ClimateSmartAg Details: doi.org/10.1016/j.pl...
New two-stage method using PointNeXt + Quickshift++ achieves 93.32% mPrec for organ instance segmentation across monocots & dicots—sugarcane, maize, tomato—outperforming 4 SOTA methods. 🌱 #PlantPhenomics #PointCloud #AI Details: doi.org/10.1016/j.pl...
Spectral imaging + machine learning diagnose herbicide modes of action in 6 hours with 89.6% accuracy—100% by day 3. Faster, cheaper weed killer screening! 🌱🔬 #AgTech #PrecisionAg #MachineLearning #Herbicide Details: doi.org/10.1016/j.pl...
NIR spectroscopy detects fusiform rust resistance in loblolly pine with 69% accuracy—no more visual guesswork. Faster, cheaper phenotyping for healthier forests! 🌲🔬 #Forestry #PlantPathology #Spectroscopy Details: doi.org/10.1016/j.pl...
Unsupervised wood-leaf separation in tree point clouds! 🌲🍃 Our network achieves 67.6% overall accuracy without any labeled data, outperforming state-of-the-art methods. No annotation needed—just raw 3D scans. #PointCloud #DeepLearning Details: doi.org/10.1016/j.pl...
UAV remote sensing meets sugar beet breeding! 🛸 High-resolution canopy data predicts root weight (R2=0.92) and sugar content (R2=0.83) to accelerate early-season selection and yield prediction.#UAV Details: doi.org/10.34133/pla...
Unlock grape secrets with SAM! Our out-of-the-box pipeline segments 3,500 cluster images to yield 150k+ berry masks (R2=0.96). Linear regression (R2=0.87) adjusts berry visibility for precise architecture mapping. #grape Details: doi.org/10.34133/pla...
Meet Point-Line Net! 🌽🤖 This new AI model overcomes complex field backgrounds to track maize leaves and stalks with 81.5% accuracy, revolutionizing automated plant breeding. Details: doi.org/10.34133/pla...
