AI-based ear counting
ear counting
Efficiently detectable at last!
is by far one of the most time-consuming analyses in cereal production. The number of ears is an important yield-relevant parameter. If the average grain size and number are known, the expected yield can be estimated from the number of ears. The number and size differences of the ears are also an important characteristic in many trial systems. They are recorded using a standard counting frame method, in which a very small area is counted in comparison to the crop or trial element and then the total number is determined by extrapolation. Drones and smartphones offer an excellent way to automate the counting and (to a certain extent) the sizing. As soon as the ears are systematically photographed, the algorithms can provide the desired parameters by means of ear recognition. However, it must be borne in mind that manual assessment typically also takes place under the canopy, i.e. in the area that the drone cannot see.
This “concealment factor” can vary depending on the variety, crop and sowing rate. As a rule, the number is therefore underestimated by the drone. Relatively speaking, however, the values can be derived with good quality and correlated with digital assessments at the beginning of the season, such as plant number, degree of cover and tillering, thus providing an overall picture. The question therefore arises as to whether the overall picture of the digital surveys over the entire season, together with variety and process knowledge, is not the more suitable approach compared to the small sample sizes of manual counting.
Digital Phenotyping – Delivered in a Single Workflow
Digital Plant assessments
Essential as a tool!
Digital assessments play a crucial role in modern agriculture and plant research. They enable precise and efficient assessment of plants and field trials, which are essential for seed and variety development, crop protection, and fertilization strategies. Traditional methods of plant evaluation are often time-consuming and labor-intensive and can only provide limited amounts of data. This is where digitalization offers groundbreaking progress. With technologies such as drones, smartphones and advanced image processing based on artificial intelligence, each plant can be captured and assessed individually. This enables a comprehensive and precise analysis of large field areas and thus significantly improves the quality and informative value of field trials.
The use of digital assessments has several advantages: it increases efficiency by automating data collection, increases the accuracy of the data, and enables a georeferenced and detailed evaluation of test plots. These techniques can be applied to various crops and issues and open up new possibilities in plant research and agriculture. Our web-based solution for the automatic evaluation of exact and strip trials is an example of the power of digital scoring. It offers a user-friendly, integrated platform seamlessly integrated into existing workflows and enables comprehensive analyses across different locations and points in time. With digital scoring, we drive plant research and agriculture innovation to develop more sustainable and efficient cultivation strategies.
Pheno-Inspect Provides The Tool – You Do the Research!
Phenotyping
Field Trials & Agricultural Research
In field trials, whether for plant breeding, seed and variety evaluation, plant protection, or fertilization, it is essential to record the phenotypic characteristics of the plants in the field with high precision and reliability. The exact recording of these parameters is the basis for successfully evaluating the trial and the resulting advice. Field trials are typically divided into exact and strip trials. The relevant trial parameters and phenotypic characteristics are recorded using precisely coordinated and timed field and plant sampling. The counting of individual plants, the assessment of weeds, and the professional evaluation of plant diseases are incredibly time-consuming and, therefore, cost-intensive. As a result, even in the “smaller” trial plots in the exact trial, only random samples are often assessed, as there is often no staff available for a comprehensive assessment. This results in an underutilized information potential of the collected parameters. This applies particularly to strip trials, where site and other effects can only be considered to a limited extent. As a result, data collection and evaluation remain below the technical possibilities already available today.
Pheno-Inspect bietet innovative Produkte zur bildbasierten Auswertung der relevanten Parameter. Zuerst werden Bilddaten der Versuche mit einfach zu bedienenden Drohnen oder Smartphones aufgenommen. Danach erfassen die „digitalen Experten“ von Pheno-Inspect, d.h. auf künstlicher Intelligenz basierende Bildverarbeitungsalgorithmen, jede einzelne Pflanze und extrahieren relevante phänotypischen Kenngrößen für das gesamte Prüfglied. Die Bildanalyse ist vollständig georeferenziert. Die Parameter werden direkt unter Berücksichtigung von Parzellen und/oder Teilflächen aus dem Versuchsplan ausgewertet und können bequem über mehrere Standorte und Zeitpunkte statistisch aufbereitet werden. Die Bildverarbeitungstechnologie kann sehr effizient für Versuchszwecke, unabhängig von Kulturart oder Fragestellung und prinzipiell auch für Fragestellungen im gesamten Pflanzenbau verwendet werden.