Remote Control of Greenhouse Vegetable Production with Artificial Intelligence—Greenhouse Climate, Irrigation, and Crop Production


Hemming, S., de Zwart, F., Elings, A., Righini, I., & Petropoulou, A. (2019). Remote Control of Greenhouse Vegetable Production with Artificial Intelligence—Greenhouse Climate, Irrigation, and Crop Production. Sensors, 19(8), 1807. 


As the global population keeps growing, so does the demand for healthy, fresh, and accessible food. Greenhouse environments contribute an important function in providing fresh food at a high production rate while maximizing resource efficiency. The greenhouse industry has stumbled over the obstacle of finding enough skilled staff to manage these crop production systems. Modern high-tech greenhouses are equipped with process computers, and, to add more automated control, greenhouse climate and crop models have been developed. The use of Artificial Intelligence (AI) has reached breakthroughs in many areas and has not been used yet to control climate and irrigation and make crop management decisions for growing a greenhouse crop autonomously. An international challenge was conducted on autonomous greenhouses in 2018 at the high-tech research greenhouses of Wageningen University and Research in cooperation with five multi-disciplinary international teams to combine the use of modern AI algorithms and the greenhouse crop production of cucumbers cultivar “Hi-Power”. This article aims to describe the results of the teams in terms of optimizing crop yields and net profit using state-of-the-art AI algorithms for cucumber production.


The experiments were conducted in six identical greenhouse compartments equipped with standard motor movers. Cucumber seedlings were grown in rockwool substrate cubes and placed on slabs on hanging gutters. The five participating teams were: Sonoma, iGrow, deep_greens, The Croperators, and AiCu. These teams were able to remotely control the motors’ activity in their greenhouse compartment by using their own original AI algorithm, varying in design and techniques. As part of the competition, a sixth team led by Dutch growers controlled a greenhouse compartment and served as a reference. The competing teams used their original AI algorithms to regulate the climate and irrigation setpoints through a central computer and operated in the greenhouse compartments accordingly. Standard sensors set up in the greenhouse compartments continuously measured data and calculations were made based on these and digitally sent back to the teams. Teams were also allowed to install additional sensors in their compartments at the start of the experiment. Three harvest quality parameters were established: A: no defects (>375g), B: defects in shape, color, or others (300-374g), and C: less than 300g per fruit. Harvest data was measured manually by greenhouse staff and digitally sent to the teams.

The teams were judged based on three criteria: Sustainability, based on resource use efficiency (20%), Net profit, based on the number of fruits harvested per price of the fruit and category (50%), and AI algorithm, based on originality and efficacy of the algorithm (30%). Two models were used to analyze and compare the different AI algorithm approaches since the operation resulted in differences in cropping, climate, irrigation strategies, harvest yields, and resource use efficiencies. However, the combined model did not represent the presence and effects of pests and disease. The combined model was carried out to compare the calculated output as the predicted fresh cucumber yield per greenhouse compartment versus the realized cucumber yield in the same greenhouse compartment to verify the models. Additionally, model calculations were carried out applying the cropping strategy, lighting strategy, and climate strategy of other teams for each greenhouse compartment to predict the changes in yield and compare these.

Figure 2 – Scheme of data exchange (Hemming et al., 2019)

Results and Discussion

Out of all the teams, including the Dutch growers, the Sonoma team resulted in the highest production of cucumbers and consistently did this throughout the experiment. They predicted that by having a high daily light integral, they would be able to achieve a greater harvest, as they did, focusing their algorithms on this. Other teams decided to increase daily light integral and maintain a low carbon dioxide concentration (The Croperators) or have lower daily light integrals in the beginning to minimize fruit abortion (Dutch growers) which resulted in lower yields but led to a greater understanding of important production factors. All teams began the experiment with a relatively low amount of carbon dioxide, where most teams increased their concentration throughout the experiment and later diminished their dosage towards the end, the Sonoma team decided to increase their carbon dioxide concentration continuously throughout the total cropping period.

The goal of maximizing crop management is finding the optimum combination of the climate, lighting, and cropping strategies. This can be achieved by attaining the greatest number of fruits per area without hindering photosynthesis. Light, for example, is the foundation for healthy growth. For this particular experiment, having higher light integrals resulted in higher yields for the teams. As predicted with the combined models, the teams iGrow, AiCu, and growers could have possibly achieved a higher yield if they would have used the lighting strategies of either team Sonoma or The Croperators. Overall, most of the teams were able to obtain a decent production using a low amount of resources while reaching a net profit close to or better than manual commercial growers.

The goal of this experiment was to see how AI algorithms would become better at making crop management decisions and their effects on the greenhouse climate and production. At the end of the cucumber crop’s growth period, all the algorithms developed for this experiment resulted in establishing a successful greenhouse environment. As this was the first successful experiment on remotely controlling a greenhouse for cucumber production using artificial intelligence algorithms, the authors demonstrated that AI algorithms can compete, and even outperform, experienced manual growers. Artificially assisted or managed greenhouse systems could be a useful tool to grow crops where there is limited knowledge. In the future, more developments and studies in this area will be needed to make AI an alternative for trained and skilled greenhouse workers and growers, which to this day, greenhouses cannot function without.

Effects of continuous or end-of-day far-red light on tomato plant growth, morphology, light absorption, and fruit production


Kalaitzoglou, P., W. van Ieperen, J. Harbinson, M. van der Meer, S. Martinakos, K. Weerheim, C.C.S. Nicole, and L.F.M. Marcelis. 2019. Effects of continuous or end-of-day far-red light on tomato plant growth, morphology, light absorption, and fruit production. Frontiers in Plant Science 10: 322


LEDs are becoming increasingly common in modern controlled environment horticultural systems. The potential energy-savings and flexibility of LEDs makes them attractive lighting options compared to traditional high-pressure sodium or metal halide bulbs. Questions remain regarding the effects of light wavelengths emitted by LEDs on plant growth. Plants respond to differing wavelengths of light by changing physical characteristics and growth. Shading causes a low red light (R) to far-red light (FR) ratio (R:FR), resulting in low levels of the active form of a key plant photoreceptor, phytochrome. This low phytochrome stationary state (PSS) leads to a variety of shade avoidance responses that affect plant morphology and development. In contrast, LEDs used for greenhouse lighting often emit low levels or zero far-red light, causing plants to have higher PSS values than sunlight. Little is known about the effects of changing R:FR ratios on photosynthesis and plant growth in greenhouses using LEDs. Using tomato as a model crop, researchers at Wageningen University investigated how tomato morphology changes in response to higher than sunlight R:FR ratio supplied by LEDs and what effect these changes have on plant light absorption and growth (Kalaitzoglou et al., 2019). In addition, the researchers were interested if a short end-of-day FR treatment (EOD-FR) could compensate for any negative effects of growing plants without FR light during daytime.


The researchers conducted two experiments (EXP1 & EXP2). In both experiments, greenhouse chambers were divided into 15 equal compartments, each containing 20 tomato plants. Each compartment was illuminated by a combination of red (95%) and blue (5%) LEDs that supplied approximately 150 mmol
m-2 s-1 of photosynthetically active radiation (PAR) over the course of a 16 hour day. Additional FR LEDs were installed to provide five different treatments based on FR intensities. In four treatments, both FR LEDs and red/blue LEDs were on at the same time during the day, resulting in plant PSS values of 0.70, 0.73, 0.80, 0.88. The fifth treatment was a 15 minute end of day FR cycle (EOD-FR) following the end of the photoperiod. To investigate the interaction between the effects of FR and solar radiation on plant morphology the first experiment (EXP1) used a blackout screen to block incoming sunlight, while the second experiment (EXP2) exposed the plants to broadband solar radiation from morning to afternoon. In addition, EXP1 lasted only four weeks after transplanting while EXP2 was extended to 16 weeks to allow for tomato fruit development.

In both experiments, researchers measured plant growth and morphology traits such as plant height, petiole angle, and leaf area. In EXP2, fruit traits including total fruit weight (g/plant), number of fruits, and number of open flowers at 4 weeks were recorded. Additional measured traits included leaf ligh absorbance and chlorophyll and carotenoid content. To simulate light absorption for each treatment, researchers constructed a 3D plant model using GroIMP software (Hemmerling et al., 2008). Researchers used the model to estimate the effects of changes in plant morphology in response to FR light treatments on plant light absorption.

Supplemental Figure S5. (Kalaitzoglou et al., 2019)


Increasing R:FR ratio to levels above sunlight had a negative impact on tomato plant growth. In both experiments, morphological parameters such as plant height and leaf area decreased as PSS values became higher (Supplemental Figure S5). The researchers concluded that lower levels of plant light absorption in high PSS treatments were primarily caused by the decrease in leaf area, ultimately reducing plant growth. Similar results were observed for fruit characteristics in which fruit size and fruit number were greater in treatments with increasing FR compared to the 0.88 PSS treatment (no FR) and EOD-FR treatment. FR treatments also stimulated early flower and fruit maturity. Interestingly, while leaf PAR absorbance and chlorophyll content were lower in low PSS treatments, net photosynthesis was higher. Researchers attributed this result to the Emerson effect in which a higher rate of photosynthesis occurs when plants are exposed to a simultaneous mixture of red and far-red light. In both studies, EOD-FR treatments were not enough to offset the negative effects of growing plants with low levels or zero FR light throughout the day. In conclusion, the results of the study indicated that the presence of FR light increased fruit yield and accelerated flowering and FR LEDs could be a beneficial addition to greenhouses to improve tomato fruit production.

Design for an Improved Temperature Integration Concept in Greenhouse Cultivation

Original paper: O. Ko¨rner *, H. Challa Design for an improved temperature integration concept in greenhouse cultivation Farm Technology Group, Department of Agrotechnology and Food Sciences, Wageningen University, Mansholtlaan 10, 6708 PA Wageningen, The Netherlands Received 22 July 2002; received in revised form 15 November 2002; accepted 28 December 2002.

     Heating energy represents more than two-thirds of a typical greenhouse total energy consumption. Is currently well known that the average day and night temperatures are what controls how fast plants develop. As the temperature increases, crops develop much faster, but there is a significant cost associated, which typically leads to an increase in energy consumption. To mitigate the cost and become more efficient in the control environment production, an approach to improve temperature integration concept could play an essential role in energy savings. Temperature Integration Concept is based on the ability of crops to tolerate temperature deviation from their biological set points. The integration concept manipulates temperature, aiming to be compensated within a pre-set period without having adverse effects on plant growth.

     Theoretically, a crop with more dynamic and flexible temperature boundaries could potentially play an important role, so this study aimed to improve the temperature integration concept by introducing dynamic temperature constraints. A modified temperature integration procedure was designed combining the usual long-term temperature average over several days and fixed boundaries for daily average temperature with short-term temperature averages over 24 hours with a very flexible temperature limit. The overall idea is based on a concept called the Freedom for temperature fluctuations. This concept allows the temperature to freely fluctuate due to the environment without being controlled by heating or ventilation. Temperature fluctuation increases with longer averaging period and increasing temperature bandwidth, which allows longer periods of several days, which enables compensation of warm or cold periods resulting in higher energy savings.

     The proposed regimen for temperature integration was performed by modeling and simulation techniques (MATLAB version 6.0) using tomato as a model crop. Variables such as air temperature, outside radiation, relative humidity, and CO2 measurements were input in the model with a fixed time of 5 min over one year. Measurements such as setpoints for heating, ventilation and CO2 concentrations were calculated with a climate control model (CCM) which provide enough information for calculating relative humidity, air temperature, energy consumption, and natural gas consumption. For accuracy, energy loses where also consider into the model. Two reference temperature regimens were used for comparison:  BP= commercial standards, setpoint increase linearly, and a Bpfix= night and daytime heating and ventilation temperature setpoints were fixed (uncommon practice). The heating setpoints were 18, and 19 °C and ventilation set points were 19 and 20 °C for night and day, respectively. The weather prediction was also used for providing data into the model simulation. Validations of the CCM model was performed in four semi-commercial Venlo-type greenhouse compartments.

     Two temperature integration regimens were model by the Autor: RTI (regular temperature integration) and MTI (modified temperature integration) both with a bandwidth of +/- 2, +/- 4 and +/- 6. The modified regime model (MTI) resulted in more energy saved when compared with regular temperature integration model (RTI) and the BP controls. Energy-saving increased with temperature bandwidth in all cases evaluated. Fluctuation during a cold time (winter) was observed. Overall, yearly greenhouse energy saving increased by up to 23% compared with the BP regime (temperature with a bandwidth of +/- 6 C). Compared with regular temperature integration energy-saving increased relatively with 14%. Interestingly, the setpoint for relative humidity profoundly influenced energy-saving suggesting further focus in future evaluations. When evaluating the different temperature dose-response data, they observe than an increase in the duration of maximum and minimum temperature increase energy saving and gross photosynthesis of tomato plants, which can be traduced to more photosynthetic efficiency. In conclusion, the conceptual design for advance temperature integration control seems to be promising for energy reduction. The distinction between short- and long-term processes in temperature integration lead to an increase in energy savings. A more advanced flexible humidity control concept could probably help to decrease energy consumption further since the highest energy saving was achieved when no humidity control was used.

Maximizing crop photosynthesis across the entire canopy requires the optimization of many environmental factors

Original paper: Körner, O., Heuvelink, E., and Niu, Q. 2009. Quantification of temperature, CO2, and light effects on crop photosynthesis as a basis for model-based greenhouse climate control. The Journal of Horticultural Science and Biotechnology. 84:233-239.


Photosynthesis is impacted by multiple environmental factors including temperature, light intensity, and carbon dioxide (CO2) concentration. If optimal environmental conditions that maximize photosynthesis are quantified, they can be employed in controlled environments to increase crop productivity. Attempts to measure such optimal conditions have been undertaken in the past. Environmental setpoint measurements from these studies have even been compiled and implemented into various mathematical models known as “crop photosynthesis models” (CPMs) that can predict potential photosynthetic activity based on a plant’s environment. However, many of the environmental setpoints used in CPMs have relied on leaf-level photosynthesis measurements and optimization which are not always compatible with canopy-wide photosynthesis optimization. This potential incompatibility is caused by differences in the microclimate between the various levels in a crop’s canopy. For example, light intensity generally decreases as you move from the top of the canopy down to lower leaves. Also, there can be large variations in individual leaf temperature and humidity throughout the canopy which will affect photosynthesis. Other studies have investigated canopy-wide photosynthesis, but many were performed in poorly-sealed greenhouses where conditions could potentially fluctuate. Oliver Körner and his colleagues sought to more accurately quantify optimal environmental conditions for canopy-wide photosynthesis by using well-sealed greenhouses equipped with air conditioning and CO2 supplementation. These environmental control measures allowed for experiments in which temperature and CO2 concentration could be effectively manipulated and accurately maintained. The ability to control CO2 concentration and measure CO2 consumption in the greenhouse system was critical to this study. Photosynthesis was quantified by monitoring the amount of CO2 consumed by the plants in the greenhouse. Minimizing any gas exchange with the natural environment was crucial to ensure any measured CO2 change was a result of photosynthesis.

The photosynthetic responses of two different crops (cut-chrysanthemum and tomato) were quantified under different temperatures and CO2 concentrations. ‘Reagan Improved’ chrysanthemum plants were exposed to different combinations of three temperature setpoints (23, 28, and 33 °C) and three CO2 concentrations (400, 700, and 1000 µmol CO2 mol-1) under natural light levels. Similarly, CO2 consumption was measured in ‘Moneymaker’ tomatoes under different combinations of three temperature setpoints (20, 26, and 32 °C) and two CO2 concentrations (400 and 1000 µmol CO2 mol-1). Increasing CO2 concentration raised the maximum potential photosynthetic rate in both crops across all tested temperature setpoints, and this effect was greater in chrysanthemum than tomato. Additionally, higher CO2 levels led to a higher photochemical efficiency (µmol CO2 µmol photons-1) in both chrysanthemum and tomato. Temperature effect on photosynthetic rate was more complicated although photochemical efficiency in both crops consistently decreased as temperature increased. In chrysanthemum and tomato, both light intensity and CO2 concentration affected how temperature affected maximum photosynthetic rate. Using discrete light intensities (600, 900, and 1200 µmol m-2 s-1), optimum temperatures for maximum photosynthesis at 400 and 1000 µmol CO2 mol-1 were calculated. In chrysanthemum, the optimum temperature at all three light intensities was below 23 °C at 400 µmol CO2 mol-1 so a trend was not clear. At the same CO2 concentration in tomatoes, optimum temperature for tomato photosynthesis increased with higher light levels, and the largest increase in optimum temperature occurred between 900 and 1200 µmol m-2 s-1 (25.3 to 27.1 °C). Optimum temperature for chrysanthemum and tomato photosynthesis at 1000 µmol CO2 mol-1 both increased when light intensities increased. At this CO2 concentration, optimum temperature changed the most in both crops when light intensity was changed from 600 to 900 µmol m-2 s-1. Specifically, chrysanthemum optimum temperature changed from 23.5 °C to 26.9 °C while tomato optimum temperature increased from 26.6 °C 28.4 °C.

Körner and his colleagues sought to quantify the optimum environmental conditions (temperature, CO2 concentration, and light intensity) for canopy-level photosynthesis in two crops (cut-chrysanthemum and tomato). Higher CO2 levels increased maximum photosynthesis and photochemical efficiency in both crops with this effect being greater at higher temperatures. Similarly, higher CO2 concentration led to an increased optimum temperature for photosynthesis, and this occurred at the largest level when light intensity was high. Variability in the canopy microclimate (most notably temperature and light intensity) resulted in different environmental factor effects than those observed in leaf-level photosynthesis models. In general, environmental conditions caused smaller changes in canopy-level photosynthesis when compared to leaf-level photosynthesis. While basic trends were similar in both chrysanthemum and tomato, the results indicate that optimum environmental conditions for photosynthesis must be quantified for individual crops. Differences between crops including leaf area and canopy architecture must be accounted for to create accurate CPMs. In conclusion, this study indicates that crop-specific responses to interactions between multiple environmental factors must be accounted for in CPMs to accurately quantify canopy-level photosynthesis.