1. Introduction
The advancement of human society is currently undergoing a revolution in information and intelligence. Big data analysis is rapidly emerging as a new scientific research model, following experimental science, theoretical analysis, and computer simulation. Crop science research is transitioning from theoretical, experimental, and computational sciences to a phase where data-intensive knowledge discovery is the research paradigm, propelling crop science into the era of big data-centered omics research [
1]. Since the 1980s, the development of genomics and high-throughput sequencing technology has ushered life sciences into a new era of omics research characterized by large samples, big data, and expansive scientific exploration.
With the vast amount of crop genome information available, it has become a new challenge to understand the interaction mechanisms between gene function, environmental response, and crop phenotype with high resolution and efficiency [
2]. Against this background, crop phenomics [
3], [
4] has developed rapidly in recent years and has become a hot area of crop science research [
5], [
6], [
7], [
8], [
9], [
10]. The strategic frontier in agricultural and life sciences now involves breakthroughs in high-throughput, automated, and precise sensing and acquisition of plant phenotypes throughout the life cycle, even under complex growth conditions. Additionally, solutions for intelligent extraction of multimodal, multi-scale, and multi-environment big data in multi-omics through cross-disciplinary collaborative research in life sciences, information sciences, and engineering sciences are crucial.
Crop phenotypes encompass physical, physiological, and biochemical attributes that manifest the structural and functional characteristics of crops at the cellular, tissue, organ, plant, and population levels. The essence of the crop phenotype is the temporal and three-dimensional (3D) manifestation of crop gene maps, coupled with geographically distinct features and intergenerational evolutionary patterns [
11]. Crop phenotype refers to the selective characterization of genetic information based on gene-environment interactions, encompassing data on inheritance, variation, and environmental response at various scales and modalities. The generation and formation of big data in crop phenomics (BDCP) is a vast scientific project due to the intricacy of crop phenotypic traits (
Fig. 1). Focused on the practical requirements of crop breeding and cultivation, BDCP utilizes a blend of agronomy, automation, mechanical engineering, remote sensing, bioinformatics, and graphic imaging, among others. BDCP strives for the multi-scale, automated collection of phenotypes throughout the life cycle, expanding upon new physical, chemical, and biological sensors, coupled with innovative phenotype acquisition technology and platform equipment, grounded in artificial intelligence and Internet of Things technology. It integrates big data algorithms, data mining, analysis, and management technology to create a comprehensive and systematic process for phenotypic big data that can be utilized for intelligent breeding design and smart cultivation.
At present, crop phenomics, deeply integrated with information technology and intelligent equipment, is pivotal for gaining an advantage in the future development of the agricultural industry. The research and development (R&D) of technology and equipment of BDCP is moving towards ① expanding into larger construction areas; ② transitioning from analyzing single-scale phenotypes to integrated, intelligent continuous monitoring throughout the entire life cycle; and ③ integrating phenomics with collaborative innovation in multi-omics. Single and individual phenotypic information is insufficient for omics association analysis in the current era. Therefore, the comprehensive and systematic gathering of phenotypic information is crucial for future research.
To acquire multi-year, multi-regional BDCP that fulfills the requirements of crop breeding and cultivation, the development of user-friendly, efficient (high accuracy and effectiveness), and economical (low cost) BDCP technology and equipment for crop science researchers is necessary. Consequently, BDCP technology and equipment hold strong potential as the next-generation digital infrastructure for the advancement and modernization of crop science.
Under the new scientific paradigm of “artificial intelligence + big data”, crop science research faces new challenges in crop phenomics [
12]. The foremost question is how to acquire high-throughput and high-quality BDCP. Another important query is what crop phenotypic data will suffice for constructing foundational models for crop phenotyping. To achieve the objective of digital transformation and upgrading, these queries need to be addressed effectively. Furthermore, the relationship between humans and crops in traditional agricultural practices is transitioning towards a crop-human-machine mode of interconnectivity, interoperability, and interdependency. As a result, researchers in the field should strive to create technologies and equipment for BDCP to address these challenges.
As R&D in crop phenomics and related industries continue to advance, establishing standards to regulate the R&D of technologies and equipment has become a pressing concern. This will ensure the high-quality development of these fields and promote their continued growth and success.
This paper aims to address the lack of comprehensive standards in crop phenomics research. We present an industrial mapping of crop phenomics and analyze the necessity, current status, and goals of establishing a standard framework. Additionally, we discuss the organizational structure of the standard framework and its key technical points. Finally, we provide insights into the prospective development of technology and equipment for BDCP.
2. Industrial mapping of technology and equipment for BDCP
2.1. Industrial mapping
Industry mapping facilitates the analysis and presentation of information and relationships within the crop phenomics industry chain. Therefore, to comprehend the entire BDCP process, we propose creating an industry map of the technology and equipment used (
Fig. 2). This will help align the construction of the standard framework with the trends in crop phenomics industry development.
The crop phenotyping industry chain comprises various key players, including users, research institutes, phenotyping companies, government agencies, and suppliers. Users fall into diverse categories such as breeding enterprises, a variety of regional test stations, research institutes, and large-scale farms. They specify the application scenarios and propose their requirements to research organizations and phenotyping companies. Research institutions receive funding from governments, the private sector, or non-governmental research funding agencies to conduct R&D on essential technologies and equipment tailored to user needs. Additionally, governments provide feedback on the processes involved in national research funding systems. This feedback can be gathered through consultation processes to define priorities or from the results of research projects, which then inform funding schemes or even legislation. This aspect is particularly important for topics such as political issues in agriculture, regulations on data access, or the field of artificial intelligence.
The R&D on essential technologies and equipment involves the integration of mature components, technologies, and devices from other fields. Subsequently, prototypes of phenotyping techniques and equipment are manufactured and iteratively developed into marketable products, with support from phenotyping companies. Research institutions and companies specializing in phenotyping offer technical assistance and services to customers, who, in turn, provide feedback to their suppliers on issues encountered when using phenotyping products. Customers also receive guidance on the optimal use of phenotyping products as a result of their feedback. As part of this process, suppliers furnish research organizations and phenotyping companies with raw materials, including sensors, and provide production and processing services.
Specifically,
Fig. 2 lists the typical users in the crop phenomics industry chain, the major research institutes, the key technologies of crop phenomics, the representative products available in the market, the companies specializing in crop phenomics, the main services provided by the crop phenomics industry, and the evaluation system (competitiveness of products and services) of the crop phenomics technology, equipment, and products. It should be noted that the listed contents do not encompass all outstanding institutes, companies, platforms, and products in plant phenomics, but represent our perspective.
2.2. Analysis of industrial problems
Crop phenomics research, development, and the application of technology and equipment for BDCP have driven the rapid growth of crop phenotype-related industries. Nevertheless, the rapid advancement of technology and equipment for BDCP also encounters numerous challenges, including:
(1) The R&D of crop phenotyping technology and equipment is mainly conducted by colleges and research institutes. While it can address particular requirements, there is a lack of universality, low reusability, and significant duplication of the R&D process [
13]. This makes it challenging to produce effective and affordable phenotyping products [
14].
(2) Crop phenotyping sensors, imaging boxes, phenotyping platforms, and analyzing software are not fully developed by a single team, resulting in inefficient upstream and downstream integration of multiple components and technologies. Consequently, the development of highly stable phenotyping products is challenging and requires a great deal of effort.
(3) Breeding enterprises and research institutions have collected copious amounts of crop phenotyping data. However, the inconsistency in data standards and individual usage makes sharing information a challenge, and data protection or legal hurdles further hinder data reuse, thus limiting the data's potential value.
(4) The limited phenotypic data structuring necessitates extensive manual interaction during phenotypic data processing, making it challenging to create pipelined and automated phenotype parsing algorithms and software capable of batch processing. This impedes the transformation of BDCP into phenotypic traits that crop cultivation and breeding specialists can use.
In summary, the primary cause of these problems is the absence of standardized crop phenomics specifications. To move from research to application, it is crucial to construct a standard framework of technology and equipment for BDCP.
3. The necessity, current status, and objectives of building the standard framework
3.1. The necessity of building the standard framework
As a next-generation digital infrastructure for advancing crop science to data science, the technology and equipment of BDCP will merge holistically with crop science. Thus, there is an urgent need to establish systematic standards to guide and regulate the high-quality development of industries related to BDCP.
From a macro perspective, there are three primary requirements for establishing a technology and equipment standard framework for BDCP:
(1) As mentioned previously, the industrial map of crop phenotyping technology and equipment has gradually become clear, resulting in the formation of a new industry. It is essential to establish relevant standards to guide industry development, phenotyping equipment, and infrastructure R&D, and improve the capacity for service delivery.
(2) Currently, there is a pressing need for high-quality BDCP in crop multi-omics research [
15], smart agriculture research and application [
16], and the emerging scientific paradigm of AI for Science in the field of agriculture. Obtaining high-quality BDCP requires adherence to standardized specifications.
(3) The digital breeding [
17] process and acceleration of breeding [
18] require managing and analyzing crop phenotyping big data, as well as implementing a socialized service model to promote industry growth. Crop phenotyping standards must guide and regulate all aspects of these processes.
3.2. Current status of standards for crop phenomics
Genomes, transcriptomes, and metabolomes have evolved over time, leading to standardized specifications like ENCODE and Expression Atlas. Crop phenomics began later than other omics, and researchers in this field have been advocating for the establishment of relevant standards in recent years to promote the industry’s rapid development. However, current standardization mainly focuses on data storage and management.
Researchers have extended the Findable, Accessible, Interoperable, Reusable (FAIR) data principles [
19] to plant phenotyping [
20], examining the advantages and challenges of its implementation in specific cases. The Minimal Information About a Plant Phenotyping Experiment (MIAPPE) standard [
13] is utilized by diverse organizations and researchers to organize, utilize, and share phenotyping data. It primarily aims for objective and clear dissemination of data and has been updated by experts in the field [
21], [
22]. Additionally, researchers have investigated issues related to the exchange and storage of phenotypic data. Although the published phenotype data has certain reusability, its scarcity [
23], limited coverage of plant species, and low completeness of supporting data make it difficult to further enhance data optimization or integration through deep learning algorithms to improve usability.
However, due to differing national conditions, levels of agricultural informatization, intelligent equipment technology, and certain phenotypes exhibiting high susceptibilities to geographical environments or other specific factors, it is challenging to directly replicate phenotyping standards among different countries. Therefore, it is necessary to formulate relevant standards according to the specific circumstances of each country.
Additionally, current standards primarily concentrate on data storage and management, insufficient to cover the entire crop phenotyping industrial chain. These standards are typically tailored for research organizations or specific high-throughput phenotyping platforms, lacking widespread implementation in both scientific research and industry.
In reality, BDCP is a comprehensive procedure that spans from designing experimental scenarios and phenotyping equipment to data acquisition and processing, data application, and standardization of crop phenomics. It is crucial to develop a standardized system that encompasses the entire crop phenotyping process rather than just concentrating on one or a few aspects. Thus, it is imperative to explore research pertaining to constructing a technology and equipment standard framework for BDCP.
3.3. The objectives of building the standard framework
The life cycle of crops and their associated scenarios in agricultural production are complex. The morphology and structure of crops are intricate, and dense canopies in field production cause severe cross-obscuration. The multi-dimensional phenotypic traits of crops are influenced by several non-structural factors, including light, temperature, humidity, and airflow, which have a significant impact. Achieving unmanned industrialized agriculture is challenging due to the nature of production.
The standard framework for technology and equipment of BDCP aims to enhance the coherence and consistency of the crop phenotyping process by applying standards. The ultimate objective is to obtain structured and reliable BDCP with good integrity. This will enable interconnection between hardware equipment, infrastructure, phenotyping algorithms, data management, and sharing, leading to the high-quality development of the crop phenotyping industry. These goals could not be achieved simultaneously and are thus the long-term objectives for the framework.
Based on industrial demand, product form, user clarity, and clear application scenarios, the following four specific short-term objectives have to be realized through the construction of a series of standards for BDCP technology and equipment (
Fig. 3):
(1) Clarify the R&D interfaces for crop phenotyping sensors, platforms, and other equipment, improve the consistency and reusability of the hardware and equipment, and thus improve the stability of the platforms and reduce the cost of production.
(2) Improve work efficiency and reduce operation and maintenance costs of phenotyping products by defining specific application scenarios.
(3) Reduce data noise and acquisition and use costs by standardizing the phenotyping data collection process, improving the structuring and consistency of phenotyping data, and improving phenotyping data. By limiting the specific application scenarios of phenotyping products, the efficiency of phenotyping products can be improved and the operation and maintenance costs can be reduced;
(4) By enhancing the degree of phenotyping data structure, provide input data of high quality and high consistency for phenotyping algorithms. By improving the degree of structured phenotypic data, we can provide high-quality and high-consistency input data for phenotyping algorithms, thus promoting the pipelining and automation level of phenotypic data processing and parsing.
4. Construction route of technology and equipment standard framework for BDCP
Based on our practical experience and reflection, this section presents the standard framework’s organizational structure and discusses crucial phases during construction.
4.1. Suggested organizational structure of the standard framework
Crops exhibit significant diversity, encompassing abundant varieties and growing under a range of different environmental conditions. Throughout their various stages of growth, crops undergo notable phenotypic variations in both morphology and structure. These factors collectively contribute to challenges in establishing consistent standards for crop cultivation. Therefore, this paper focuses on the requirements and features of the crop phenotyping sector. It extracts content with a considerable degree of commonality to establish a standardized framework while deprioritizing content that exhibits high degrees of individualization and is challenging to standardize at present.
Fig. 4 displays the technological and equipment standard framework of BDCP, comprising three primary components: development standards for crop phenotyping hardware and equipment, acquisition standards for crop phenotyping data, and specifications for the storage and management of crop phenotyping data. In the crop phenotyping industry, the hardware equipment for acquiring phenotypes, data collection methods, and other processes in the previous stages significantly influence subsequent data processing, analysis, and application. Consequently, they receive higher priority and dominate most of the content within the standard framework. While phenotype extraction and analysis algorithms rely on standardized data, they must also accommodate the specific application needs of required phenotypic traits, necessitating a high degree of personalization. As a result, establishing a universally applicable standard specification is challenging and is not included in the standard framework.
Crop phenotyping hardware equipment R&D standards consist of three main components: sensors, imaging boxes, and phenotyping platform R&D standards. Considering that environmental factors such as light, wind, and rainfall significantly affect crop phenotype acquisition in the field, while indoor environments offer greater control, crop phenotyping data acquisition standards are divided into two sections: field and controlled environments. The crop phenotype data storage management standard focuses on the format of both raw and analyzed data, as well as the completeness of supporting information. It does not include subdivisions.
The technology and equipment standards of BDCP operate within a comprehensive system of interdependent relationships and cannot function independently. For example, the data acquisition specifications of the rail-based phenotyping platform in the field are directly influenced by the mounted imaging box of the platform. When considering subsequent crop phenotype storage and management specifications, the format of the acquired data and other related issues should also be taken into account. It is essential to note that during standard development, finer details among organizational sub-nodes must be regulated. This can be achieved by subdividing into multiple standard specifications, including those for different digital breeding and intelligent cultivation application scenarios, as well as other subdivisions.
4.2. Key points of standard framework construction
In recent years, we have carried out a lot of work around the research and development of crop phenotyping platforms in multi-environments [
24], [
25], [
26], [
27], [
28], [
29], phenotyping data acquisition [
30], [
31], [
32], [
33], and multi-scale crop phenotypes extraction algorithms and software development [
34], [
35], [
36], [
37], and have accumulated rich experience. On this basis, some detailed suggestions for the construction of the standard framework are given.
4.2.1. Essentials for the construction of crop phenotyping hardware equipment R&D standards
The crop phenotyping hardware and equipment development standard aims to standardize and guide the research and development of electronic components, equipment, infrastructure, and technological systems for crop phenotyping. It targets product development and service providers, including suppliers, research institutes, and companies in the crop phenotyping industry map.
(1) Sensor R&D/integration standards. Crop phenotyping sensors are primarily dominated by integrated mature sensors; however, they are not customized or tailored for crop phenotyping research and development. For instance, light detection and ranging (LiDAR) is mainly intended for research and development of long-distance tasks such as autopilot and target detection, resulting in low point cloud resolution in nearby terrestrial phenotyping platform scenarios. Similarly, red, green, and blue (RGB) cameras are primarily intended for industrial imaging or consumer-grade tasks in R&D, while the focal length and stability of field scenes in phenotyping tasks require considerable improvement. In crop phenotyping, upgrading or replacing the sensor for selection presents multiple obstacles encompassing the size, load-bearing, communication, control, and other related issues of the platform docking system. Additionally, variances in sensor output data format, resolution, and other disparities bear significant implications for the ensuing phenotyping analysis algorithm. Therefore, the research and development or integration of crop phenotyping sensors is a crucial component of the crop phenotyping standard framework. Crop phenotyping sensors comprise RGB cameras, multispectral sensors, hyperspectral sensors, thermal imaging cameras, LiDAR, ion sensors, and small molecule sensors. Standards for R&D or sensor integration should prioritize aspects such as sensor size, weight, effective distance, communication interface protocols, control mode, data resolution, data acquisition interval, and data output format.
(2) Imaging box R&D standards. The crop phenotyping imaging box is typically mounted onto a transmission device, creating a phenotyping platform. The specific needs of different indicators require various platform types, such as those mounted on unmanned aerial vehicles (UAVs), robotics, and unmanned ground vehicles (UGVs). These platforms require smaller sizes, lighter weights, and lower power consumption. In contrast, rail-based or pipelined phenotyping platforms require imaging boxes that meet tighter limitations in terms of size, weight, and power consumption. These platforms are more concerned with data resolution and stability. Crop phenotyping imaging systems should prioritize factors such as size, weight, power usage, interface protocols, data output formats, control methods, data resolution, and other relevant indicators.
(3) Phenotyping platform R&D standards. Crop phenotyping platforms, akin to agricultural machinery, come in various styles and require consideration of the combination of information, agronomy, and machinery. Therefore, standards are necessary to regulate the development of phenotyping platform products. UAVs, rail-based systems, UGVs, robots, pipelined devices, and portable equipment constitute crop phenotyping platforms due to differences in application scenarios, objectives, construction costs, and operation and maintenance expenses. The imaging and movement modes of phenotyping platforms are directly influenced by ground and soil conditions, as well as light and wind speeds within field and controlled environments. Moreover, differences in construction, operation, and maintenance further categorize platforms into two types: those designed for field use and those for facility use. Specifically, for the standards governing the development and construction of phenotyping platforms, it is crucial for rail-based phenotyping platforms to prioritize ground reinforcement to prevent prolonged usage that could lead to platform collapse or rail deformation. Additionally, the platform’s height range and x−y axis motion range should be considered. The standards focus on the operational speed of the imaging box, as well as other indicators such as remote control, path design, and acquisition of time and location coordinates. The two types of rail-based platforms, designed for both field and controlled environments, may have different performance criteria. Rail-based phenotyping platforms in the field must avoid construction in windy locations. Pipelined phenotyping platforms should prioritize indicators such as plant load bearing, maximum plant size, and ground-to-height range. UGVs and robotic phenotyping platforms should prioritize their maximum sensor or imaging box capacity, load-bearing capacity, power consumption, and endurance time. To mitigate data quality issues caused by unstable terrain and bumps, the platforms' standards should incorporate stability indicators. UGV phenotyping platforms should consider factors such as body length, platform elevation, and steering method. The development of robotic phenotype platforms should align with production scenarios in typical fields or controlled environments. It is necessary to limit the body’s maximum width so that it can navigate between crop rows. The platform must have functions for path planning, autonomous navigation, and obstacle avoidance. Additionally, it should have the ability to perform path planning, autonomous navigation, and obstacle avoidance.
4.2.2. Essentials for the construction of crop phenotyping data collection standards
Unlike crop phenotyping hardware and equipment development, which primarily targets front-end groups in the industrial chain, crop phenotyping data acquisition standards focus mainly on the back end of the industrial chain, specifically users of phenotyping products. As a result, they have a broader audience and stronger binding force, crucial for ensuring that superior phenotyping products can reach their full potential. By establishing standards for data acquisition and guiding users to obtain high-quality data, crop phenotyping standards can provide structured and complete inputs for subsequent phenotype extraction algorithms and applications, ultimately enhancing the automation of these processes.
(1) Essentials for phenotyping data acquisition standard in the field.
The field environment is more complex than controlled environments, necessitating the development of data acquisition standards to improve data consistency. The acquisition must be limited to specific times and weather conditions; this includes setting maximum wind speed limits during data collection and refraining from collecting spectral data during cloudy weather. Clouds can cause significant fluctuations in solar radiation, resulting in inaccurate data. Relevant calibrations must be performed for every data acquisition, including color calibration of RGB images and spectrum calibration for spectrum images.
Different phenotyping platforms focus on various data acquisition standard points. For UAV phenotyping platforms, flight altitudes should be standardized based on specific requirements. Additionally, the overlap between adjacent routes and the maximum angle of aerial footage should be determined. For rail-based phenotyping platforms, specifications should include sensor or imaging box height from the ground, timing of data acquisition, and overlap between adjacent paths. When it comes to UGV and robotics phenotyping platforms, it’s crucial to consider ground conditions when entering planting areas to prevent harm to crops. All data obtained from mobile phenotyping platforms should be accompanied by time and location information, enabling linkage with plots or plants during subsequent phenotyping analysis. Furthermore, since the field environment’s light, temperature, humidity, and other parameters fluctuate rapidly, crop phenotype-environmental data synchronization standards should be included to enhance data accessibility.
(2) Essentials for phenotyping data acquisition standard in controlled environments.
Facilities and indoor environments offer greater stability, and crops should be planted in accordance with the requirements of the phenotyping platform, facilitating data acquisition while simultaneously meeting research and production goals. It is imperative to maintain cleanliness in controlled environments, such as eliminating crop residue from conveyor belts of pipelined phenotyping platforms, and the imaging room, to minimize the impact of noise. Indoor micro-scale crop phenotyping [
30] necessitates sample preparation, accounting for limiting factors such as sampling sites and sample moisture content, to ensure consistent data.
Furthermore, it is necessary for every data acquisition protocol, whether in outdoor or indoor settings, to encompass preliminary sensor and platform evaluations, data acquisition documentation, and data validation to guarantee reliable data.
4.2.3. Essentials for the construction of crop phenotyping data storage and management standards
Research on the phenotypic identification of crop germplasm resources, the formation mechanism of important crop traits, and the demand for accelerated breeding necessitate multi-year, multi-regional experiments to acquire crop phenotypic data. Effectively storing, managing, and sharing phenotypic data is crucial for circulating and increasing the value of such information. This includes addressing environmental diversity, variation in varieties and cultivation methods, and ensuring consistent standards of phenotypic data acquisition. This remains a fundamental purpose of present international phenotyping networks (e.g., International Plant Phenotyping Network (IPPN), European Plant Phenotyping Network (EPPN), China Plant Phenotyping Network (CPPN), etc.).
By employing sensors and phenotyping platforms, BDCP can be acquired in compliance with data acquisition standards. However, BDCP faces challenges of inadequate data oversight and challenging data circulation, resulting in significant data silos, low data quality, and making it difficult to leverage the potential of data elements fully. A unified standard is required for storing and managing such large datasets to achieve data sharing, retrieval, and application.
Relevant standards for phenotypic data management have been developed [
13], and a brief discussion is presented here in accordance with the proposed standard framework. The storage and management standard for BDCP emphasizes data format, quality, and integrity. The data format should determine the formatting of RGB images, point clouds, and spectral data to enable their use without requiring conversion by users. It is essential to establish relevant standards to ensure explicit data quality, limit high-quality data levels, and assess compliance with data acquisition standards. Additionally, data completeness requires the inclusion of essential details such as plant varieties, cultivation and management procedures, environmental meteorology, acquisition platforms, acquisition personnel, and data acquisition timelines. Some information is preserved for future reference. Among other considerations, data trust models should be studied to promote data protection and data reuse. The BDCP storage and management standard can assist in developing a crop phenotyping database [
38], [
39] and promoting the creation of a repository for sharing and utilizing BDCP data, akin to Github, enabling genuine sharing and application of BDCP. The well-organized BDCP bears the potential of being combined with related genetic and environmental information [
40] for enhancing the proficiency of knowledge retrieval via multi-omics analysis and exploration.
A checklist (
Table 1) summarizes the principal elements of the framework, including hardware, data acquisition, data storage, and organization. This would serve as a reference for calls for proposals, evaluations, projects, or data management plans.
5. Concluding remarks and future perspectives
5.1. Promotion of the standard framework for the entire process of crop phenotyping
The primary aim of crop phenomics is to procure accurate and dependable BDCP. Illustrated in
Fig. 5 are the sequential steps involved in generating and implementing BDCP. Initially, data acquisition technologies such as sensors, imaging units, and phenotyping platforms are developed to collect the initial set of phenotypic data, encompassing RGB and spectral images, 3D point clouds of crops, and traditional manual measurements. Subsequently, this data undergoes analysis through phenotyping algorithms, including image segmentation, point cloud alignment, and deep learning techniques. Through this process, various phenotypic traits pertaining to morphology, structure, color, texture, physiology, biochemistry, and fertility dynamics are derived, forming the foundation of a comprehensive crop phenotyping database. These traits serve multiple purposes, ranging from identifying germplasm resources to conducting multi-omics correlation analysis and facilitating field production management decisions. However, it is essential to acknowledge that the life cycle of these phenotypic data is critical to realizing the full potential of BDCP. Phenotypic data of suboptimal quality, or those left unanalyzed or underutilized, significantly diminish their intended value.
The development of technology and equipment for BDCP, coupled with the establishment of crop phenotyping standards, aims to maximize the exploitation and utilization of BDCP’s value. The construction of the crop phenotyping standard framework has four pivotal effects on the overall crop phenotyping process (refer to
Fig. 5): ① It can help standardize and guide the research and development of crop phenotyping equipment, thereby enhancing process efficiency and stability of the phenotyping process. ② Additionally, it can standardize and guide the process of obtaining crop phenotyping data to better the consistency of initial phenotyping data and ensure data quality. ③ It can provide structured and high-quality data for phenotyping algorithms, improving their processing capability and automation level. ④ Additionally, it standardizes and guides the construction of crop phenome databases, promoting the storage, management, sharing, and application of phenome big data.
High throughput and accuracy are the two key aspects of phenotyping. Achieving both simultaneously in a single process presents a challenge. However, the phenotyping standards framework standardizes the entire process, encompassing both the hardware of the phenotyping platform to enhance throughput and the quality of data acquisition to improve accuracy. This objective evaluation highlights its potential benefits for phenotyping research and crop improvement. Moreover, the utilization of high-quality and extensive BDCP can facilitate the construction of pre-trained artificial intelligence models, advancing the development of general artificial intelligence within the agricultural industry. Overall, the advancement of technology and equipment standards for BDCP will not only facilitate the progress of crop phenomics as a discipline but also cultivate the growth of the crop phenotyping industry, indirectly supporting the advancement of research and application of digital breeding and intelligent cultivation.
5.2. Evaluation methods for the standard framework
The objective of establishing a technology and equipment standard framework for BDCP is to enhance the development of the crop phenotyping industry. Hence, the method of developing the framework should consider what constitutes superior phenotyping products. In our view, an effective phenotyping product should be both user-friendly and cost-effective, designed to meet the needs of main users such as agricultural university graduates or specialists. It should have the capacity for high-throughput, automation, and batch processing, while truly addressing the challenges of collecting phenotyping information in agricultural research and production. Therefore, when developing standards, it is essential to consider the specific needs and circumstances of each country to establish a crop phenotyping standard framework that aligns with national conditions.
When formulating standards for crop phenotyping, it is important to consider evaluating their effectiveness. This includes assessing whether the standards and their corresponding systems are successful or flawed. The assessment of the standard framework at the hardware level primarily hinges on the ease of integration among the sensors, imaging boxes, and phenotyping platforms. At the data acquisition level, the quality of the data serves as the main criterion for framework evaluation. The assessment of the standard framework at the data storage, management, and application level largely depends on the framework’s ability to apply and circulate the collected data. Most importantly, the phenotyping standard system must be validated. This implies that it must be accepted by the majority of those involved in phenotyping research, product development, and users. Ultimately, the most important evaluator is whether it facilitates the healthy development of phenomics and the industry.
5.3. Recommendations for the standard framework construction
Crop phenomics is a multidisciplinary field spanning agronomy, mechanical engineering, automation, bioinformatics, remote sensing, graphic imaging, and related domains. Hence, formulating a technology and equipment standard framework for BDCP requires collaboration among multidisciplinary research and development professionals. Additionally, it’s crucial to ensure compatibility with established standards in other fields like industrial sensors, agricultural equipment, 5th generation mobile communication technology (5G), and big data management and storage. However, the complexity of the agricultural sector means that the crop phenotyping standard framework must prioritize certain aspects over universality. Once the overarching framework is established, it can be further tailored for specific applications in crop breeding and cultivation.
The development of the standard framework for technology and equipment for BDCP, along with setting standard specifications, is a nuanced process that demands continuous communication between multidisciplinary R&D professionals and end-users. This involves creating standards for the framework and articulating the needs of the crop phenomics industry, while also refining and improving the standards over time. Given the multitude of components within the standard framework, it's prudent to focus on formulating standards for the front end of the industrial chain initially, based on strong demand and attention. These standards can then be progressively enhanced and expanded. Additionally, the establishment of phenotyping standards in different countries should be carefully considered, with international phenotyping organizations aiming to harmonize standards across borders during implementation.
The development and application of technology and equipment for BDCP can be greatly aided by the implementation of a standard framework for crop phenotyping, which acts as a catalyst for the industry. This leads to the efficient functioning of BDCP technology and equipment, enhances the governance of BDCP assets, and drives advancements in crop breeding and smart cultivation. Moreover, it empowers value-added practices in agriculture.
Acknowledgments
This work was partially supported by the National Key R&D Program of China (2022YFD2002300), the Construction of Collaborative Innovation Center of Beijing Academy of Agricultural and Forestry Sciences (KJCX20240406), the National Natural Science Foundation of China (32071891), and the earmarked fund (CARS-02 and CARS-054).
Compliance with ethics guidelines
Weiliang Wen, Shenghao Gu, Ying Zhang, Wanneng Yang, and Xinyu Guo declare that they have no conflict of interest or financial conflicts to disclose.
PDF
(3885KB)
