Image Processing Workflow Research Articles

Multi-species cryoEM calibration and workflow verification standard.

Cryogenic electron microscopy (cryoEM) is a rapidly growing structural biology modality that has been successful in revealing molecular details of biological systems. However, unlike established biophysical and analytical techniques with calibration standards, cryoEM has lacked comprehensive biological test samples. Here, a cryoEM calibration sample consisting of a mixture of compatible macromolecules is introduced that can not only be used for resolution optimization, but also provides multiple reference points for evaluating instrument performance, data quality and image-processing workflows in a single experiment. This combined test specimen provides researchers with a reference point for validating their cryoEM pipeline, benchmarking their methodologies and testing new algorithms.

Improved Brain Tumor Detection MRI Using Advanced Processing Techniques: Enhancement and Convolution Case Studies

Brain tumors present a significant challenge in medical imaging due to their complexity, requiring early detection and precise analysis for effective treatment. This study develops and evaluates advanced image processing workflows aimed at enhancing brain tumor image analysis. The proposed method involves four main steps: enlargement, pre-processing with min-max filters, enhancement, and convolution. The dataset used is from Kaggle, comprising 3,364 images categorized into Glioma (100 images), Meningioma (115 images), No Tumor (105 images), and Pituitary Tumor (74 images). For this study, images from the Glioma, Meningioma, and Pituitary Tumor categories were used, with one image selected from each category for technique evaluation. The results showed significant improvements in image clarity and detail, with high correlation values of 0.9851 for Meningioma and 0.9886 for Pituitary. These findings highlight the effectiveness of the proposed techniques in enhancing image quality and diagnostic accuracy.

Setting up an institutional OMERO environment for bioimage data: Perspectives from both facility staff and users.

Modern bioimaging core facilities at research institutions are essential for managing and maintaining high-end instruments, providing training and support for researchers in experimental design, image acquisition and data analysis. An important task for these facilities is the professional management of complex multidimensional bioimaging data, which are often produced in large quantity and very different file formats. This article details the process that led to successfully implementing the OME Remote Objects system (OMERO) for bioimage-specific research data management (RDM) at the Core Facility Cellular Imaging (CFCI) at the Technische Universität Dresden (TU Dresden). Ensuring compliance with the FAIR (findable, accessible, interoperable, reusable) principles, we outline here the challenges that we faced in adapting data handling and storage to a new RDM system. These challenges included the introduction of a standardised group-specific naming convention, metadata curation with tagging and Key-Value pairs, and integration of existing image processing workflows. By sharing our experiences, this article aims to provide insights and recommendations for both individual researchers and educational institutions intending to implement OMERO as a management system for bioimaging data. We showcase how tailored decisions and structured approaches lead to successful outcomes in RDM practices.

BrainSuite BIDS App: Containerized Workflows for MRI Analysis.

There has been a concerted effort by the neuroimaging community to establish standards for computational methods for data analysis that promote reproducibility and portability. In particular, the Brain Imaging Data Structure (BIDS) specifies a standard for storing imaging data, and the related BIDS App methodology provides a standard for implementing containerized processing environments that include all necessary dependencies to process BIDS datasets using image processing workflows. We present the BrainSuite BIDS App, which encapsulates the core MRI processing functionality of BrainSuite within the BIDS App framework. Specifically, the BrainSuite BIDS App implements a participant-level workflow comprising three pipelines and a corresponding set of group-level analysis workflows for processing the participant-level outputs. The Anatomical Pipeline extracts cortical surface models from a T1-weighted (T1w) MRI. It then performs surface-constrained volumetric registration to align the T1w MRI to a labeled anatomical atlas, which is used to delineate anatomical regions of interest in the MRI brain volume and on the cortical surface models. The Diffusion Pipeline processes diffusion-weighted imaging (DWI) data, with steps that include coregistering the DWI data to the T1w scan, correcting for susceptibility-induced geometric image distortion, and fitting diffusion models to the DWI data. The Functional Pipeline performs fMRI processing using a combination of FSL, AFNI, and BrainSuite tools. It coregisters the fMRI data to the T1w image, then transforms the data to the anatomical atlas space and to the Human Connectome Project's grayordinate space. The outputs of each pipeline can then be processed during group-level analysis. The outputs of the Anatomical Pipeline and the Diffusion Pipeline are analyzed using the BrainSuite Statistics Toolbox in R (bstr), which provides functionality for hypothesis testing and statistical modeling. The outputs of the Functional Pipeline can be analyzed using atlas-based or atlas-free statistical methods during group-level processing. These analyses include the application of BrainSync, which synchronizes the time-series data temporally and enables comparison of resting-state or task-based fMRI data across scans. We also present the BrainSuite Dashboard quality control system, which provides a browser-based interface for reviewing the outputs of individual modules of the participant-level pipelines across a study in real-time as they are generated. BrainSuite Dashboard facilitates rapid review of intermediate results, enabling users to identify processing errors and make adjustments to processing parameters if necessary. The comprehensive functionality included in the BrainSuite BIDS App provides a mechanism for rapidly deploying the BrainSuite workflows into new environments to perform large-scale studies. We demonstrate the capabilities of the BrainSuite BIDS App using structural, diffusion, and functional MRI data from the Amsterdam Open MRI Collection's Population Imaging of Psychology dataset.

Multi-species cryoEM calibration and workflow verification standard.

Cryogenic electron microscopy (cryoEM) is a rapidly growing structural biology modality that has been successful in revealing molecular details of biological systems. However, unlike established biophysical and analytical techniques with calibration standards, cryoEM has lacked comprehensive biological test samples. We introduce a cryoEM calibration sample that is a mixture of compatible macromolecules that can be used not only for resolution optimization but also provides multiple reference points for evaluating instrument performance, data quality, and image processing workflows in a single experiment. This combined test specimen provides researchers a reference point for validating their cryoEM pipeline, benchmarking their methodologies, and testing new algorithms.

Rapid Microbial Profiling through Multimodal Biosensors for Transversal Analysis.

The intricate interactions between host and microbial communities hold significant implications for biology and medicine. However, traditional microbial profiling methods face limitations in processing time, measurement of absolute abundance, detection of low biomass, discrimination between live and dead cells, and functional analysis. This study introduces a rapid multimodal microbial characterization platform, Multimodal Biosensors for Transversal Analysis (MBioTA), for capturing the taxonomy, viability, and functional genes of the microbiota. The platform incorporates single cell biosensors, scalable microwell arrays, and automated image processing for rapid transversal analysis in as few as 2 h. The multimodal biosensors simultaneously characterize the taxon, viability, and functional gene expression of individual cells. By automating the image processing workflow, the single cell analysis techniques enable the quantification of bacteria with sensitivity down to 0.0075%, showcasing its capability in detecting low biomass samples. We illustrate the applicability of the MBioTA platform through the transversal analysis of the gut microbiota composition, viability, and functionality in a familial Alzheimer's disease mouse model. The effectiveness, rapid turnaround, and scalability of the MBioTA platform will facilitate its application from basic research to clinical diagnostics, potentially revolutionizing our understanding and management of diseases associated with microbe-host interactions.

Deep-Learning-Based Automatic Sinkhole Recognition: Application to the Eastern Dead Sea

Sinkholes can cause significant damage to infrastructures, agriculture, and endanger lives in active karst regions like the Dead Sea’s eastern shore at Ghor Al-Haditha. The common sinkhole mapping methods often require costly high-resolution data and manual, time-consuming expert analysis. This study introduces an efficient deep learning model designed to improve sinkhole mapping using accessible satellite imagery, which could enhance management practices related to sinkholes and other geohazards in evaporite karst regions. The developed AI system is centered around the U-Net architecture. The model was initially trained on a high-resolution drone dataset (0.1 m GSD, phase I), covering 250 sinkhole instances. Subsequently, it was additionally fine-tuned on a larger dataset from a Pleiades Neo satellite image (0.3 m GSD, phase II) with 1038 instances. The training process involved an automated image-processing workflow and strategic layer freezing and unfreezing to adapt the model to different input scales and resolutions. We show the usefulness of initial layer features learned on drone data, for the coarser, more readily-available satellite inputs. The validation revealed high detection accuracy for sinkholes, with phase I achieving a recall of 96.79% and an F1 score of 97.08%, and phase II reaching a recall of 92.06% and an F1 score of 91.23%. These results confirm the model’s accuracy and its capability to maintain high performance across varying resolutions. Our findings highlight the potential of using RGB visual bands for sinkhole detection across different karst environments. This approach provides a scalable, cost-effective solution for continuous mapping, monitoring, and risk mitigation related to sinkhole hazards. The developed system is not limited only to sinkholes however, and can be naturally extended to other geohazards as well. Moreover, since it currently uses U-Net as a backbone, the system can be extended to incorporate super-resolution techniques, leveraging U-Net based latent diffusion models to address the smaller-scale, ambiguous geo-structures that are often found in geoscientific data.

STOPGAP: an open-source package for template matching, subtomogram alignment and classification.

Cryo-electron tomography (cryo-ET) enables molecular-resolution 3D imaging of complex biological specimens such as viral particles, cellular sections and, in some cases, whole cells. This enables the structural characterization of molecules in their near-native environments, without the need for purification or separation, thereby preserving biological information such as conformational states and spatial relationships between different molecular species. Subtomogram averaging is an image-processing workflow that allows users to leverage cryo-ET data to identify and localize target molecules, determine high-resolution structures of repeating molecular species and classify different conformational states. Here, STOPGAP, an open-source package for subtomogram averaging that is designed to provide users with fine control over each of these steps, is described. In providing detailed descriptions of the image-processing algorithms that STOPGAP uses, this manuscript is also intended to serve as a technical resource to users as well as for further community-driven software development.

Spatiospectral image processing workflow considerations for advanced MR spectroscopy of the brain.

Magnetic resonance spectroscopy (MRS) is one of the few non-invasive imaging modalities capable of making neurochemical and metabolic measurements in vivo. Traditionally, the clinical utility of MRS has been narrow. The most common use has been the "single-voxel spectroscopy" variant to discern the presence of a lactate peak in the spectra in one location in the brain, typically to evaluate for ischemia in neonates. Thus, the reduction of rich spectral data to a binary variable has not classically necessitated much signal processing. However, scanners have become more powerful and MRS sequences more advanced, increasing data complexity and adding 2 to 3 spatial dimensions in addition to the spectral one. The result is a spatially- and spectrally-variant MRS image ripe for image processing innovation. Despite this potential, the logistics for robustly accessing and manipulating MRS data across different scanners, data formats, and software standards remain unclear. Thus, as research into MRS advances, there is a clear need to better characterize its image processing considerations to facilitate innovation from scientists and engineers. Building on established neuroimaging standards, we describe a framework for manipulating these images that generalizes to the voxel, spectral, and metabolite level across space and multiple imaging sites while integrating with LCModel, a widely used quantitative MRS peak-fitting platform. In doing so, we provide examples to demonstrate the advantages of such a workflow in relation to recent publications and with new data. Overall, we hope our characterizations will lower the barrier of entry to MRS processing for neuroimaging researchers.

Three-Dimensional Characterization of Dry Particle Coating Structures Originating from the Mechano-fusion Process.

Dry particle coating processes are of key importance for creating functionalized materials. By a change in surface structure, initiated during coating, a surface property change and thus functionalization can be achieved. This study introduces an innovative approach employing 3D X-ray micro-computed tomography (micro-CT) to characterize coated particles, consisting of spherical alumina particles (d50 = 45.64 μm), called hosts, surrounded by spherical polystyrene particles (d50 = 3.5 μm), called guests. The formed structures, hetero-aggregates, are generated by dry particle coating using mechano-fusion (MF). A deeper understanding of the influence of MF process parameters on the coating structures is a crucial step toward tailoring of coating structure, resulting surface property and functionalization. Therefore, the influence of rotational speed, process time, and total mechanical energy input during MF is explored. Leveraging micro-CT data, acquired of coated particles, enables non-stereologically biased and quantitative coating structure analysis. The guest's coating thickness is analyzed using the maximum inscribed sphere and ray method, two different local thickness measurement approaches. Particle-discrete information of the coating structure are available after a proper image processing workflow is implemented. Coating efficiency and guest's neighboring relations (nearest neighbor distance and number of neighbors inside search radius) are evaluated.

Deep Learning-Based Psoriasis Assessment: Harnessing Clinical Trial Imaging for Accurate Psoriasis Area Severity Index Prediction

Introduction: Image-based machine learning holds great promise for facilitating clinical care; however, the datasets often used for model training differ from the interventional clinical trial-based findings frequently used to inform treatment guidelines. Here, we draw on longitudinal imaging of psoriasis patients undergoing treatment in the Ultima 2 clinical trial (NCT02684357), including 2,700 body images with psoriasis area severity index (PASI) annotations by uniformly trained dermatologists. Methods: An image-processing workflow integrating clinical photos of multiple body regions into one model pipeline was developed, which we refer to as the “One-Step PASI” framework due to its simultaneous body detection, lesion detection, and lesion severity classification. Group-stratified cross-validation was performed with 145 deep convolutional neural network models combined in an ensemble learning architecture. Results: The highest-performing model demonstrated a mean absolute error of 3.3, Lin’s concordance correlation coefficient of 0.86, and Pearson correlation coefficient of 0.90 across a wide range of PASI scores comprising disease classifications of clear skin, mild, and moderate-to-severe disease. Within-person, time-series analysis of model performance demonstrated that PASI predictions closely tracked the trajectory of physician scores from severe to clear skin without systematically over- or underestimating PASI scores or percent changes from baseline. Conclusion: This study demonstrates the potential of image processing and deep learning to translate otherwise inaccessible clinical trial data into accurate, extensible machine learning models to assess therapeutic efficacy.

A service-based approach to cryoEM facility processing pipelines at eBIC.

Electron cryo-microscopy image-processing workflows are typically composed of elements that may, broadly speaking, be categorized as high-throughput workloads which transition to high-performance workloads as preprocessed data are aggregated. The high-throughput elements are of particular importance in the context of live processing, where an optimal response is highly coupled to the temporal profile of the data collection. In other words, each movie should be processed as quickly as possible at the earliest opportunity. The high level of disconnected parallelization in the high-throughput problem directly allows a completely scalable solution across a distributed computer system, with the only technical obstacle being an efficient and reliable implementation. The cloud computing frameworks primarily developed for the deployment of high-availability web applications provide an environment with a number of appealing features for such high-throughput processing tasks. Here, an implementation of an early-stage processing pipeline for electron cryotomography experiments using a service-based architecture deployed on a Kubernetes cluster is discussed in order to demonstrate the benefits of this approach and how it may be extended to scenarios of considerably increased complexity.

Sizing mudsnails: Applying superpixels to scale growth detection under ocean warming

Abstract The expansion of scientific image data holds great promise to quantify individuals, size distributions and traits. Computer vision tools are especially powerful to automate data mining of images and thus have been applied widely across studies in aquatic and terrestrial ecology. Yet marine benthic communities, especially infauna, remain understudied despite their dominance of marine biomass, biodiversity and playing critical roles in ecosystem functioning. Here, we disaggregated infauna from sediment cores taken throughout the spring transition (April–June) from a near‐natural mesocosm setup under experimental warming (Ambient, +1.5°C, +3.0°C). Numerically abundant mudsnails were imaged in batches under stereomicroscopy, from which we automatically counted and sized individuals using a superpixel‐based segmentation algorithm. Our segmentation approach was based on clustering superpixels, which naturally partition images by low‐level properties (e.g., colour, shape and edges) and allow instance‐based segmentation to extract all individuals from each image. We demonstrate high accuracy and precision for counting and sizing individuals, through a procedure that is robust to the number of individuals per image (5–65) and to size ranges spanning an order of magnitude (<750 μm to 7.4 mm). The segmentation routine provided at least a fivefold increase in efficiency compared with manual measurements. Scaling this approach to a larger dataset tallied >40k individuals and revealed overall growth in response to springtime warming. We illustrate that image processing and segmentation workflows can be built upon existing open‐access R packages, underlining the potential for wider adoption of computer vision tools among ecologists. The image‐based approach also generated reproducible data products that, alongside our scripts, we have made freely available. This work reinforces the need for next‐generation monitoring of benthic communities, especially infauna, which can display differential responses to average warming.

A web-based dashboard for RELION metadata visualization.

Cryo-electron microscopy (cryo-EM) has witnessed radical progress in the past decade, driven by developments in hardware and software. While current software packages include processing pipelines that simplify the image-processing workflow, they do not prioritize the in-depth analysis of crucial metadata, limiting troubleshooting for challenging data sets. The widely used RELION software package lacks a graphical native representation of the underlying metadata. Here, two web-based tools are introduced: relion_live.py, which offers real-time feedback on data collection, aiding swift decision-making during data acquisition, and relion_analyse.py, a graphical interface to represent RELION projects by plotting essential metadata including interactive data filtration and analysis. A useful script for estimating ice thickness and data quality during movie pre-processing is also presented. These tools empower researchers to analyse data efficiently and allow informed decisions during data collection and processing.

Overcoming resolution attenuation during tilted cryo-EM data collection

Structural biology efforts using cryogenic electron microscopy are frequently stifled by specimens adopting “preferred orientations” on grids, leading to anisotropic map resolution and impeding structure determination. Tilting the specimen stage during data collection is a generalizable solution but has historically led to substantial resolution attenuation. Here, we develop updated data collection and image processing workflows and demonstrate, using multiple specimens, that resolution attenuation is negligible or significantly reduced across tilt angles. Reconstructions with and without the stage tilted as high as 60° are virtually indistinguishable. These strategies allowed the reconstruction to 3 Å resolution of a bacterial RNA polymerase with preferred orientation, containing an unnatural nucleotide for studying novel base pair recognition. Furthermore, we present a quantitative framework that allows cryo-EM practitioners to define an optimal tilt angle during data acquisition. These results reinforce the utility of employing stage tilt for data collection and provide quantitative metrics to obtain isotropic maps.

Virtual Orientation Tools (VOTj): Fiji plugins for object centering and alignment.

Standardizing image datasets is essential for facilitating overall visual comparisons and enhancing compatibility with image-processing workflows. One way to achieve homogeneity for images containing a single object is to align the object to a common orientation. Here, we propose the Virtual Orientation Tools (VOTj): a set of Fiji plugins to center and align an object of interest in images to a vertical or horizontal orientation. To process an image, the plugin requires either a mask outlining the object or a rough annotation of the object directly drawn by the user in the image. The current object orientation is retrieved using Principal Component Analysis (PCA), from which the optimal alignment is derived. The plugins support multi-dimensional images to allow, e.g., aligning individual time points of a time-lapse. The tools can be used for a variety of samples and imaging modalities. Besides, the plugins enable the interactive alignment of a list of images from a directory for batch execution and can be included in custom image-processing workflows using macro-recording.

Research on the application of machine vision in the design of machine automation

Abstract The application of machine vision technology to mechanical automation design is an inevitable measure to improve the comprehensive productivity of mechanical production manufacturing. In this paper, a mechanical automation system based on SOA architecture is designed to cover the image acquisition, processing, and recognition workflow of an image processing module based on machine vision. The binocular camera calibration is completed by the Zhang Zhengyou calibration method. The image is preprocessed by using algorithms such as the mean value method, segmented linear transformation and median filtering, and an algorithm is proposed to improve the ORB feature point extraction, which can extract and match the feature points quickly and efficiently. Meanwhile, a Harris corner point detection algorithm is proposed to improve the SIFT algorithm to enhance the accuracy of target recognition and localization. The designed mechanical automation system is applied to the lychee picking robot, for example, analysis, which shows that the overall recognition P-R value of the system reaches 0.953, and the classification accuracy is above 0.917. In the mechanical automation route localization, the lateral deviation is lower than 4.80cm. The maximum time for image processing and parameter transfer is 117.966ms, which indicates that the system involved in this paper is relatively stable in operation and has a better effect on the localization of the navigation line, which is of certain application value in the field of mechanical automation design.

EARLY STAGES OF COLORECTAL CANCER CHARACTERIZATION BY AUTOFLUORESCENCE 3D MICROSCOPY: A PRELIMINARY STUDY.

Colorectal cancer is one of the most prevalent pathologies worldwide whose prognosis is linked to early detection. Colonoscopy is the gold standard for screening, and diagnosis is usually made histologically from biopsies. Aiming to reduce the inspection and diagnostic time as well as the biopsies and resources involved, other techniques are being promoted to conduct accurate in vivo colonoscopy assessments. Optical biopsy aims to detect normal and neoplastic tissues analysing the autofluorescence spectrum based on the changes in the distribution and concentration of autofluorescent molecules caused by colorectal cancer. Therefore, the autofluorescence contribution analysed by image processing techniques could be an approach to a faster characterization of the target tissue. Quantify intensity parameters through digital processing of two data sets of three-dimensional widefield autofluorescence microscopy images, acquired by fresh colon tissue samples from a colorectal cancer murine model. Additionally, analyse the autofluorescence data to provide a characterization over a volume of approximately 50 µm of the colon mucosa for each image, at second (2nd), fourth (4th) and eighth (8th) weeks after colorectal cancer induction. Development of a colorectal cancer murine model using azoxymethane/dextran sodium sulphate induction, and data sets acquisition of Z-stack images by widefield autofluorescence microscopy, from control and colorectal cancer induced animals. Pre-processing steps of intensity value adjustments followed by quantification and characterization procedures using image processing workflow automation by Fiji's macros, and statistical data analysis. The effectiveness of the colorectal cancer induction model was corroborated by a histological assessment to correlate and validate the link between histological and autofluorescence changes. The image digital processing methodology proposed was then performed on the three-dimensional images from control mice and from the 2nd, 4th, and 8th weeks after colorectal cancer chemical induction, for each data set. Statistical analyses found significant differences in the mean, standard deviation, and minimum parameters between control samples and those of the 2nd week after induction with respect to the 4th week of the first experimental study. This suggests that the characteristics of colorectal cancer can be detected after the 2nd week post-induction. The use of autofluorescence still exhibits levels of variability that prevent greater systematization of the data obtained during the progression of colorectal cancer. However, these preliminary outcomes could be considered an approach to the three-dimensional characterization of the autofluorescence of colorectal tissue, describing the autofluorescence features of samples coming from dysplasia to colorectal cancer. • A new digital image processing method was developed to measure intensity in 3D autofluorescence images of colorectal samples using a CRC mouse model. • This method showed that autofluorescence intensity in colon mucosa is similar in healthy tissue but changes significantly in tumor development. • Statistical analysis revealed CRC traits detectable from the second week post-induction, aiding in early CRC detection. • The study provides a basis for 3D autofluorescence characterization in colorectal tissue from dysplasia to cancer, although variability in autofluorescence limits data systematization during cancer progression.

Empirical Modelling Workflow for Resolution Invariant Assessment of Osteophytes.

Traditional quantitative analysis of bone microstructure in micro-computed tomography (microCT) is dependent on animal scale and requires parametric tuning in new implementations. This study aims to develop an automated and resolution-invariant 3D image processing workflow for quantitative assessment of osteophytes. In this workflow, cortical bone was segmented from microCT scans, and a 3D sphere-fitting transform was performed to obtain a thickness map, for which each voxel is assigned a thickness value corresponding to the size of the largest sphere containing the voxel that fits entirely within the cortical bone. From the thickness map, a 1-voxel thick outer surface was extracted to model surface roughness. The thickness values of the outer surface were empirically estimated by a series of known statistical distributions. Resulting parameters describing best-fit distributions, along with other cortical bone metrics, were analysed to determine sensitivity to osteoarthritis and the presence of osteophytes. The workflow was validated using microCT scans and histological gradings of rabbit and rat tibiofemoral joints. Visual inspection shows that samples with osteoarthritis and the presence of osteophytes have more surface voxels assigned small thickness values. The distribution of surface thickness values for each animal is best described by Gamma distributions, whose shape parameter is consistently sensitive to osteoarthritis and the presence of osteophytes. Combining traditional image processing with empirical distribution fitting provides an automated, objective, and resolution-invariant workflow for osteophyte assessment. The proposed method is simple, yet elegant in its implementation, and can be readily used in new implementations.

Spinach Yield Mapping Using Multispectral UAV Imagery

Crop yield prediction and mapping are essential for crop management and decision-making. The development of unmanned aerial vehicles (UAVs) accelerates the acquisition of high-resolution imagery for the evaluation and monitoring of crop growth and development. In this study, a cost-effective yield mapping workflow was developed in a commercial setting. Two UAVs were deployed to rapidly obtain RGB and multispectral images of spinach at different growth stages in multiple spinach fields to assess crop physiological attributes in the field environment. Based on the imagery, an image processing workflow for farm field orthoimage rotation and segmentation was proposed. Then, twelve vegetation indices (VIs) were extracted from the processed images. The most robust VIs were selected by comparing the correlations among the VIs and yields and regression tree and stepwise multiple linear regression. Excess Green Index (ExG) and Normalized Difference Vegetation Index (NDVI) were identified as two of the most robust VIs for predicting spinach yield at various stages of growth. A linear regression model was developed using 45 manually harvested calibration locations and a coefficient of determination (R2) of 0.98. The root mean squared error (RMSE) was 0.19 kg/m with a mean absolute percentage error (MAPE) of 9.0 % when using a validation dataset. The produced yield maps provided the basis for a farm harvesting decision-making process and for a general crop management strategy.