Smart Horticulture: Latest Advances and Prospects

1. Introduction

The latest Hunger Hotspots Report by the Food and Agriculture Organisation (FAO) and the World Food Programme (WFP) of the United Nations highlighted 16 hotspots in which food insecurity is increasing due to conflict, economic instability, and extreme weather events [1]. Moreover, the global population is growing by approximately 1.1% per year, and the majority of the world’s population is concentrated in a few low-income countries [2]. Providing sufficient and affordable food for the global population remains challenging in such a scenario. Concurrently, governments are expected to ensure food security without harming the climate, biodiversity, or natural resources [3,4].

Interest in smart technologies is steadily increasing in the horticultural sector [5], enabling real-time monitoring and fast decision-making. The employment of smart agriculture is the key concept of innovative farming. Technologies based on artificial intelligence (AI), big data, and the Internet of Things (IoT) hold significant promise in terms of increasing agricultural yield and reducing production inputs [6]. Horticulture could benefit from the progress being made in sensing technologies, robotics, and AI by implementing new labour-and cost-saving approaches [7].

This Special Issue “Smart Horticulture: Latest Advances and Prospects”, aims to provide a global perspective on the opportunities and constraints of smart horticulture. The most recent research exploring technologies and approaches applicable in both controlled environments and open fields is included.

The collection of articles included in this Special Issue underlines the relevance of an interdisciplinary approach to ensure a constant update of smart horticulture.

This Special Issue also fills relevant gaps in the current knowledge and provides a basis for future research aimed at efficient and sustainable horticultural practises.

2. Overview of Published Articles

This Special Issue includes 11 research articles which explore three main topics: (i) advances in yield monitoring and mapping, (ii) smart horticulture for crop status assessment, and (iii) technologies for process implementation. The contributions involve several countries: Brazil, Egypt, Saudi Arabia, the United Arab Emirates, Australia, Turkey, Romania, Croatia, Taiwan, Latvia, Sweden, Serbia, Japan, the UK, Malaysia, and Italy.

2.1. Advances in Yield Monitoring and Mapping

The contributions to yield monitoring concern different crops, such as coffee (Coffea arabica L.), radish (Raphanus sativus L. var. sativus), Japanese quince (Chaenomeles japonica (Thumb.)), and fruits in general. The high precision of a custom yield monitor mounted on a harvester and connected to a computer for automatically registering data allowed the yield variability within coffee fields to be captured (Contribution 1). Equipping coffee harvesters with yield monitors can potentially improve site-specific management strategies.

Random forest models based on radish colour, shape, and volume obtained from RGB images captured from multiple directions suggest the potential for the non-destructive monitoring of radish weight (Contribution 2).

A relevant contribution to the topic is given by a review investigating 189 patents of sensors for measuring fruit properties (Contribution 3). To date, the majority of sensors focus on the detection of individual fruit parameters. However, much effort has been dedicated to developing sensors detecting complex parameters, often coupled with artificial intelligence tools.

The attempt to evaluate the length and width of Japanese quince with 3D imaging provided different levels of effectiveness over eleven genotypes (Contribution 4). The study suggests that phenotyping needs to be improved, in terms of imaging processing and machine learning models.

2.2. Smart Horticulture for Crop Status Assessment

Stress detection is one of the most relevant topics analysed in this Special Issue. Three new spectral vegetation indices calculated from the reflectance of squash (Cucurbita pepo L.) leaves were effective in distinguishing water stress from potassium deficiency (Contribution 5). Similarly, an effective early detection model for drought stress was built for greenhouse tomato (Solanum lycopersicum L.) (Contribution 6). Moreover, the experiment addressed the problems of collinearity of spectral data using the random forest algorithm and resampling techniques. The effective determination of Genovese basil (Ocimum basilicum L.) moisture content was obtained through NIR spectroscopy (Contribution 7).

Plant health monitoring of indoor plants is crucial for correct and effective management, and Leaf Soil–Plant Analysis Development (SPAD) prediction is an excellent health monitoring system. Analysing ten different indoor plant species, research proved that deep neural network models are more accurate than the conventional machine learning approach in predicting leaf SPAD (Contribution 8).

Finally, four essential canopy state variables (leaf area index, canopy chlorophyll content, aboveground biomass, and fractional vegetation cover) of Bambara groundnut (Vigna subterranea (L.) Verdc.) were efficiently predicted through remote-sensing-based models (Contribution 9).

2.3. Technologies for Process Implementation

Hydroponic cultivation is considered a practice that can alleviate the pressure on food security. However, managing the nutrient content in the hydroponic solution is complex. A spectroscopic sensor was successfully tested to monitor nitrogen variations in nutrient solutions for Micro Indoor Smart Hydroponics (MISH) systems (Contribution 10). The test represents the basis for implementing simple, low-cost IoT sensors for MISH.

In vitro plant propagation is vital for the mass multiplication of pathogen-free plants. Machine learning algorithms have successfully been used to predict crucial parameters, such as shoot diameter, number of siblings, and the number of main roots in black chokeberry (Aronia melanocarpa (Michx.) Elliott) produced using tissue culture techniques (Contribution 11).

3. Conclusions

The contributions in this Special Issue highlight that the path towards smart agriculture is well defined, with active research and advanced technologies available to the field. The overarching goal of achieving sustainability in horticulture while ensuring global food security is evident across the studies. The experiments published in this Special Issue contribute to the implementation of sustainable horticultural practises. However, critical next steps remain, including refining yield forecasting models, advancing imaging processes, and focusing research on major crops. Addressing these challenges will further solidify the foundation for a sustainable and food-secure future in agriculture.