A5063605405- Smart Agriculture and AI
- Video Surveillance and Tracking Methods
- Advanced Image and Video Retrieval Techniques
- Remote Sensing in Agriculture
- Spectroscopy and Chemometric Analyses
- Fuzzy Logic and Control Systems
- Leaf Properties and Growth Measurement
- Robotics and Sensor-Based Localization
- Neural Networks and Applications
- Listeria monocytogenes in Food Safety
- Image Retrieval and Classification Techniques
- Anomaly Detection Techniques and Applications
- Advanced Vision and Imaging
- Soil Moisture and Remote Sensing
- Plant Disease Management Techniques
- Artificial Immune Systems Applications
- Video Analysis and Summarization
- Geophysical and Geoelectrical Methods
- Indoor and Outdoor Localization Technologies
- Genetic Mapping and Diversity in Plants and Animals
- Soil Geostatistics and Mapping
- Plant responses to elevated CO2
- Video Coding and Compression Technologies
- Infrared Target Detection Methodologies
- Distributed Control Multi-Agent Systems
Syngenta (Switzerland)
2024
Rothamsted Research
2016-2023
Lancaster University
2010-2016
Lancaster City Council
2016
Current approaches to field phenotyping are laborious or permit the use of only a few sensors at time. In an effort overcome this, fully automated robotic platform with dedicated sensor array that may be accurately positioned in three dimensions and mounted on fixed rails has been established, facilitate continual high-throughput monitoring crop performance. Employed comprise high-resolution visible, chlorophyll fluorescence thermal infrared cameras, two hyperspectral imagers dual 3D laser...
The detection of wheat heads in plant images is an important task for estimating pertinent traits including head population density and characteristics such as health, size, maturity stage, the presence awns. Several studies have developed methods from high-resolution RGB imagery based on machine learning algorithms. However, these generally been calibrated validated limited datasets. High variability observational conditions, genotypic differences, development stages, orientation makes a...
The Global Wheat Head Detection (GWHD) dataset was created in 2020 and has assembled 193,634 labelled wheat heads from 4700 RGB images acquired various acquisition platforms 7 countries/institutions. With an associated competition hosted Kaggle, GWHD_2020 successfully attracted attention both the computer vision agricultural science communities. From this first experience, a few avenues for improvements have been identified regarding data size, head diversity, label reliability. To address...
Crop yield is an essential measure for breeders, researchers and farmers comprised of may be calculated by the number ears/m2, grains per ear thousand grain weight. Manual wheat counting, required in breeding programmes to evaluate crop potential, labour intensive expensive; thus, development a real-time head counting system would significant advancement. In this paper, we propose computationally efficient called DeepCount automatically identify count spikes digital images taken under...
Recording growth stage information is an important aspect of precision agriculture, crop breeding and phenotyping. In practice, still primarily monitored by-eye, which not only laborious time-consuming, but also subjective error-prone. The application computer vision on digital images offers a high-throughput non-invasive alternative to manual observations its use in agriculture phenotyping increasing. This paper presents automated method detect wheat heading flowering stages, uses the...
Recently, surveillance, security, patrol, search, and rescue applications increasingly require algorithms methods that can work automatically in real time. This paper reports a new real-time approach based on three novel techniques for automatic detection, object identification, tracking video streams, respectively. The novelty detection identification are the newly proposed recursive density estimation (RDE) method. RDE is using Cauchy-type of kernel, which calculated recursively as opposed...
Accurately segmenting vegetation from the background within digital images is both a fundamental and challenging task in phenotyping. The performance of traditional methods satisfactory homogeneous environments, however, decreases when applied to acquired dynamic field environments. In this paper, multi-feature learning method proposed quantify growth outdoor conditions. introduced technique compared with state-of the-art other on images. All are evaluated different environmental conditions...
Abstract Genetic studies increasingly rely on high-throughput phenotyping, but the resulting longitudinal data pose analytical challenges. We used canopy height from an automated field phenotyping platform to compare several approaches scanning for quantitative trait loci (QTLs) and performing genomic prediction in a wheat recombinant inbred line mapping population based up 26 sampled time points (TPs). detected four persistent QTLs (i.e. expressed most of growing season), with both...
Image segmentation is a fundamental but critical step for achieving automated high- throughput phenotyping. While conventional methods perform well in homogenous environments, the performance decreases when used more complex environments. This study aimed to develop fast and robust neural-network-based tool phenotype plants both field glasshouse environments high-throughput manner. Digital images of cowpea (from glasshouse) wheat field) with different nutrient supplies across their full...
Digitalisation has opened a wealth of new data opportunities by revolutionizing how images are captured. Although the cost generation is no longer major concern, management and processing have become bottleneck. Any successful visual trait system requires automated structuring retrieval model to manage, search, retrieve unstructured complex image data. This paper investigates highly scalable computationally efficient for real-time content-based searching through large-scale repositories in...
(1) Background: Information rich hyperspectral sensing, together with robust image analysis, is providing new research pathways in plant phenotyping. This combination facilitates the acquisition of spectral signatures individual organs as well detailed information about physiological status plants. Despite advances technology field-based phenotyping, little known characteristic shaded and sunlit components wheat canopies. Non-imaging sensors cannot provide spatial information; thus, they are...
Sustainable fertilizer management in precision agriculture is essential for both economic and environmental reasons. To effectively manage input, various methods are employed to monitor track plant nutrient status. One such method hyperspectral imaging, which has been on the rise recent times. It a remote sensing tool used physiological changes response conditions availability. However, conventional processing mainly focuses either spectral or spatial information of plants. This study aims...
In this paper, a new approach for Self-evolving PArameter-free fuzzy Rule-based Controller (SPARC) is proposed. Two illustrative examples are provided aiming proof of concept. The proposed controller can start with no pre-defined rules, and does not need to pre-define the range output or control variables. This SPARC learns autonomously from its own actions while performing plant. It use any parameters, explicit membership functions, off-line pre-training nor model (e.g. in form differential...
A new approach to autonomously detect and track moving objects in a video captured by camera from UAV real-time is proposed this paper. The introduced replaces the need for human operator perform analytics detecting clustering them tracking purposes. effectiveness of tested on footage taken real evaluation results are demonstrated
In this paper, an online self-evolving fuzzy controller is proposed for autonomous leader/follower. The starts with a simple configuration and learns from its own actions while controlling the mobile robot during leader following behaviour. A traditional Takagi-Sugeno type also implemented compared to verify reliability performance of controller. Experiments are carried out real Pioneer 3DX at Lancaster University.
Detection of wheat heads is an important task allowing to estimate pertinent traits including head population density and characteristics such as sanitary state, size, maturity stage the presence awns. Several studies developed methods for detection from high-resolution RGB imagery. They are based on computer vision machine learning generally calibrated validated limited datasets. However, variability in observational conditions, genotypic differences, development stages, orientation...
The need to find related images from big data streams is shared by many professionals, such as architects, engineers, designers, journalist, and ordinary people. Users quickly the relevant generated a variety of domains. challenges in image retrieval are widely recognized, research aiming address them led area content-based becoming "hot" area. In this paper, we propose novel computationally efficient approach, which provides high visual quality result based on use local recursive density...
ABSTRACT Several studies have explored the potential of electrical resistivity tomography to monitor changes in soil moisture associated with root water uptake different crops. Such usually use a set limited below‐ground measurements throughout growth season but are often unable get complete picture dynamics processes. With development high‐throughput phenotyping platforms, we now capability collect more frequent above‐ground measurements, such as canopy cover, enabling comparison data. In...
In this paper, we present a novel approach for automatic object detection and also using on-line trajectory clustering RT anomaly in video streams. The proposed is based on two main steps. the first step, recently introduced called Recursive Density Estimation (RDE) used novelty detection. This method Cauchy type of kernel which works frame-by-frame basis does not require pre-defined threshold to identify objects. second multifeature clustered anomalies To an anomaly, trajectories are...
A new approach to autonomously detect and track a moving object in video captured by camera (possibly mounted on unmanned vehicle, UxV) is proposed this paper. It based combination of the recently introduced recursive density estimation (RDE) well-known scale invariant feature transformation (SIFT). The involves building model background using RDE sequences camera. was robust many videos with absence image registration (pixel position alignment). output cluster foreground pixels which can be...
The Global Wheat Head Detection (GWHD) dataset was created in 2020 and has assembled 193,634 labelled wheat heads from 4,700 RGB images acquired various acquisition platforms 7 countries/institutions. With an associated competition hosted Kaggle, GWHD successfully attracted attention both the computer vision agricultural science communities. From this first experience 2020, a few avenues for improvements have been identified, especially perspective of data size, head diversity label...