Advances in automated image analysis using open-source computer vision tools, such as PlantCV, have greatly increased the throughput of aboveground phenotyping in a variety of crop species. However, PlantCV was largely optimized to analyze images collected under controlled laboratory conditions, and has seldom been used to analyze images collected under field conditions. Further, there are no known applications of PlantCV for analyzing images collected belowground, such as those obtained from minirhizotron imaging devices. In this study, we demonstrated applications of PlantCV for extracting plant trait information from aboveground and belowground images collected in two perennial crop mapping populations. The first population was composed of nearly 1,200 individuals of a potential perennial oilseed crop (Silphium Integrifolium x Perfoliatum ), and the second population was composed of nearly 1,700 individuals of a perennial cover crop (Trifolium ambiguum , Kura Clover). We designed and used a field-based imaging cart to collect overhead and profile images of individuals from both populations in August and October, which improved the efficiency of field-based image capture. Around the time of aboveground image collection, belowground images of root networks were collected using minirhizotron imaging devices. We then assessed the application of PlantCV for measuring aboveground traits (crop canopy area, height, leaf color and growth rates) and belowground traits (root length and growth rates), and we explored future directions of PlantCV for field-based image analysis of aboveground and belowground crop tissues.

PlantCV is an open-source open-development image analysis software package for plant phenotyping written in Python that has been actively developed since 2014. A new version of PlantCV was recently released. Major goals of the version 4 release were to 1) simplify the process of developing workflows by reducing the amount of coding needed; 2) broadening the set of supported data types; and 3) introducing interactive annotation tools that can be used directly in PlantCV workflow notebooks. Here we highlight the use of point annotations that can be used to quickly collect sets of points for parameterization of functions such as regions of interest or the identification of landmark points. Another application of point annotations this for image annotation, which is a major bottleneck in plant phenomics. For example, we have used point annotations to analyze microscopy images aimed at measurement of quinoa salt bladders, the number and size of stomata, and scoring of pollen germination. These tasks have traditionally been low throughput and have required manual scoring, but our point annotation tools can be used along with traditional segmentation methods to semi-automatically detect and annotate images. The PlantCV point annotation tools also allow users to correct semi-automated detection results before classification (e.g., germinated vs non-germinated pollen) and extraction of size & color traits per object. Once images are annotated, results can be analyzed directly or potentially can be used as labeled data in supervised learning methods.

Multi-plant imaging using arrays of low-cost cameras is a successful strategy for capturing affordable high-throughput plant phenotyping data. An imaging platform of this type can enable simultaneous imaging of hundreds to thousands of plants. The resulting datasets enable analysis of dynamic plant growth, development, and environmental responses at high temporal resolution. Full analysis of these datasets requires the identification of individual plants for measurement, but computational separation of individual plants becomes challenging when neighboring plants overlap. Here, we introduce the use of the watershed transform to segment moderately overlapping plants in multi-plant time series datasets. Rather than focusing on segmenting plants in individual images, we utilize information encoded in the entire time series to propagate plant labels from an early time point when individual plants are separate to later time points. In preliminary studies, using this method allowed us increase the analyzable size of the dataset by 28%.

High-throughput plant phenotyping is increasingly implemented in a wide array of experimentation and presents challenges both logistically and analytically. Phenotype data are often longitudinal and proper modeling of plant growth requires sophisticated modeling techniques to account for the intra-plant correlations and changing variation over time (heteroskedasticity). For this reason, plant growth is often analyzed by comparing only single time points or the start and end points for inference with no regard for the trends themselves. Single time point analysis can be sufficient for simple biological comparisons, but modeling has the potential to unlock additional insights by utilizing all the information at hand. Current plant growth modeling strategies do account for intra-plant correlations but are still limited to constant variance assumptions and therefore perform sub-optimally. Here we propose a Bayesian hierarchical approach as an alternative method for plant growth modeling by demonstrating the utility of heteroskedastic sub-model parameterizations. We show that accounting for heteroskedasticity greatly improves model accuracy and subsequent inference. Additionally, Bayesian methodologies inherently lend themselves to near real-time model updating and we propose integration with Clowder to facilitate adaptive experimental designs. We show by example the utility of Bayesian updating and how it relates to experimental decision making.

The development of new phenomics approaches to image and process data from the subcellular to ecosystem-scale has accelerated over the past decade. Many of these tools are produced “in-house” within a single lab or a group of collaborating labs, making it hard to keep up with the state-of-the-art for phenomics hardware and software development. The Plant Cell Atlas Phenomics Committee is creating a collaborative space that connects phenomics developers with each other and with the greater plant science community, with the goal of facilitating wide-reaching collaborations. To do this, we will be hosting a video series called “How We Built It” where developers provide a short tour of their inventions and relevant biological applications. We will follow the series with a more in-depth networking event where plant scientists can connect with the inventors and discuss collaborative opportunities. Our goal is to streamline the invention of new phenotyping tools and broaden the application of existing tools.