Autonomous Greenhouse Research Articles

Intelligent agricultural robotic detection system for greenhouse tomato leaf diseases using soft computing techniques and deep learning

The development of soft computing methods has had a significant influence on the subject of autonomous intelligent agriculture. This paper offers a system for autonomous greenhouse navigation that employs a fuzzy control algorithm and a deep learning-based disease classification model for tomato plants, identifying illnesses using photos of tomato leaves. The primary novelty in this study is the introduction of an upgraded Deep Convolutional Generative Adversarial Network (DCGAN) that creates augmented pictures of disease tomato leaves from original genuine samples, considerably enhancing the training dataset. To find the optimum training model, four deep learning networks (VGG19, Inception-v3, DenseNet-201, and ResNet-152) were carefully compared on a dataset of nine tomato leaf disease classes. These models have validation accuracy of 92.32%, 90.83%, 96.61%, and 97.07%, respectively, when using the original PlantVillage dataset. The system then uses an enhanced dataset with ResNet-152 network design to achieve a high accuracy of 99.69%, as compared to the original dataset with ResNet-152’s accuracy of 97.07%. This improvement indicates the use of the proposed DCGAN in improving the performance of the deep learning model for greenhouse plant monitoring and disease detection. Furthermore, the proposed approach may have a broader use in various agricultural scenarios, potentially altering the field of autonomous intelligent agriculture.

Automatic Control of a Sunlight Reflector Board for Achieving the Sunlight Intensity Set by the Greenhouse Operator

This article proposes an autonomous greenhouse system which has a sunlight Reflector Board (RB). In general, a greenhouse, which grows leafy vegetables, utilizes artificial light, such as LED light, as a supplementary source of light. However, artificial light cannot be used as the main source of light for photosynthesis for various kinds of plants. Our article utilizes natural sunlight, instead of artificial light, for growing plants inside a greenhouse. We develop an autonomous system to control the amount of sunlight inside the greenhouse by controlling the angle of the RB, which is located in front of the greenhouse. Since the proposed greenhouse uses natural sunlight, one can grow various kinds of plants which cannot grow under artificial light. Suppose that we accumulate the RB angle data for one year. Since solar motion repeats every year, we can control the RB according to this accumulated angle data. In this way, we can reduce unnecessary RB controls, after running the system for more than a year. As far as we know, the proposed greenhouse system based on autonomous RB controls is novel, since we control RBs for providing optimal sunlight for plants inside the greenhouse. We perform experiments to demonstrate the performance of the proposed greenhouse system.

Comparative analysis of mature tomato detection by feature extraction and machine learning for autonomous greenhouse robots

Accurate detection of tomatoes grown in greenhouses is important for timely harvesting. In this way, it is ensured that mature tomatoes are collected by distinguishing them from the unripe ones. Insufficient light, occlusion, and overlapping adversely affect the detection of mature tomatoes. In addition, it is time consuming for people to detect mature tomatoes at certain periods in large greenhouses. For these reasons, high-performance automatic detection of tomatoes by greenhouse robots has become an increasingly studied area today. In this paper, two feature extraction methods, histogram of oriented gradients (HOG) and local binary patterns (LBP), which are effective in object recognition, and two important and commonly used classifiers of machine learning, support vector machines (SVM) and k-nearest neighbor (kNN), are comparatively used to detect and count tomatoes. The HOG and LBP features are classified separately and together by SVM or kNN, and the success of each case are compared. Performance of the detection is improved by eliminating false positive results at the postprocessing stage using color information.

Optimal temperature and humidity control for autonomous control system based on PSO‐BP neural networks

AbstractIn order to solve the problems of difficult control, poor stability, and low control precision in complex autonomous non‐linear systems, and some sensors have non‐linear errors in special environments. Based on the PSO (Particle Swarm Optimization) algorithm, an PSO‐BP‐PID (Particle Swarm Optimization Back Propagation neural network PID) control method and a sensor error compensation algorithm based on BP (Back Propagation) neural network are designed for optimal temperature and humidity control and sensor error compensation in the autonomous greenhouse system. The error between the average temperature value and the target value after steady state is 0.5°C, and the error between the average humidity value and the target value is 1% RH. The results show that the control method can effectively compensate the non‐linear error of the sensor and improve the performance of the control system in a complex environment, which is suitable for the stable and control of actuators in autonomous systems. The error of temperature and humidity sensor is compensated by BP neural network; PSO (Particle Swarm Optimization) was used to optimize the BP‐PID parameters of the automatic greenhouse system.

Lettuce Production in Intelligent Greenhouses-3D Imaging and Computer Vision for Plant Spacing Decisions.

Recent studies indicate that food demand will increase by 35-56% over the period 2010-2050 due to population increase, economic development, and urbanization. Greenhouse systems allow for the sustainable intensification of food production with demonstrated high crop production per cultivation area. Breakthroughs in resource-efficient fresh food production merging horticultural and AI expertise take place with the international competition "Autonomous Greenhouse Challenge". This paper describes and analyzes the results of the third edition of this competition. The competition's goal is the realization of the highest net profit in fully autonomous lettuce production. Two cultivation cycles were conducted in six high-tech greenhouse compartments with operational greenhouse decision-making realized at a distance and individually by algorithms of international participating teams. Algorithms were developed based on time series sensor data of the greenhouse climate and crop images. High crop yield and quality, short growing cycles, and low use of resources such as energy for heating, electricity for artificial light, and CO2 were decisive in realizing the competition's goal. The results highlight the importance of plant spacing and the moment of harvest decisions in promoting high crop growth rates while optimizing greenhouse occupation and resource use. In this paper, images taken with depth cameras (RealSense) for each greenhouse were used by computer vision algorithms (Deepabv3+ implemented in detectron2 v0.6) in deciding optimum plant spacing and the moment of harvest. The resulting plant height and coverage could be accurately estimated with an R2 of 0.976, and a mIoU of 98.2, respectively. These two traits were used to develop a light loss and harvest indicator to support remote decision-making. The light loss indicator could be used as a decision tool for timely spacing. Several traits were combined for the harvest indicator, ultimately resulting in a fresh weight estimation with a mean absolute error of 22 g. The proposed non-invasively estimated indicators presented in this article are promising traits to be used towards full autonomation of a dynamic commercial lettuce growing environment. Computer vision algorithms act as a catalyst in remote and non-invasive sensing of crop parameters, decisive for automated, objective, standardized, and data-driven decision making. However, spectral indexes describing lettuces growth and larger datasets than the currently accessible are crucial to address existing shortcomings between academic and industrial production systems that have been encountered in this work.

Toward Autonomous Farming—A Novel Scheme Based on Learning to Prediction and Optimization for Smart Greenhouse Environment Control

The greenhouse industry has received great attention and experienced tremendous growth in the recent past across the globe. However, energy consumption and labor cost in greenhouses account for more than 50% of the cost of greenhouse production. This demands an Internet of Things (IoT)-based smart solution for automation of greenhouse environment-related activities to ensure maintenance of the desired climate inside the greenhouse to maximize plant production with optimal resource utilization. To this end, several models are proposed in the literature that is based on a selected artificial intelligence (AI) algorithm which is once trained and then deployed. The drawback of such systems is that the trained models are fixed (locked) and, therefore, unable to adapt to dynamically changing conditions, which results in performance degradation. Second, the existing studies on the subject matter are focused on the individual key component (i.e., prediction, optimization, and control). In this article, a novel scheme is presented based on the integration of the key components, and the performance of prediction and optimization components is further enhanced through the exploitation of artificial neural network (ANN)-based learning modules to support autonomous greenhouse environment monitoring and control. For experimental analysis, the greenhouse environment is emulated through the mathematical formulation of essential greenhouse processes, considering the impact of actuators’ operations and external weather conditions. Real environmental data collected for Jeju Island, South Korea is used for model validation and result analysis. Proposed learning-based optimization scheme results are compared with two other schemes, i.e., baseline scheme and optimization scheme. Comparative analysis of the results shows that the proposed model maintains the desired indoor environment for maximizing plant production with reduced energy consumption, i.e., it achieves 61.97% reduced energy consumption than the baseline scheme, 11.73% better than the optimization scheme without learning modules. Furthermore, the proposed model achieves 67.96% and 12.56% reduction in cost when compared to the baseline scheme and optimization scheme without learning modules, respectively.

Estimation of Greenhouse Lettuce Growth Indices Based on a Two-Stage CNN Using RGB-D Images.

Growth indices can quantify crop productivity and establish optimal environmental, nutritional, and irrigation control strategies. A convolutional neural network (CNN)-based model is presented for estimating various growth indices (i.e., fresh weight, dry weight, height, leaf area, and diameter) of four varieties of greenhouse lettuce using red, green, blue, and depth (RGB-D) data obtained using a stereo camera. Data from an online autonomous greenhouse challenge (Wageningen University, June 2021) were employed in this study. The data were collected using an Intel RealSense D415 camera. The developed model has a two-stage CNN architecture based on ResNet50V2 layers. The developed model provided coefficients of determination from 0.88 to 0.95, with normalized root mean square errors of 6.09%, 6.30%, 7.65%, 7.92%, and 5.62% for fresh weight, dry weight, height, diameter, and leaf area, respectively, on unknown lettuce images. Using red, green, blue (RGB) and depth data employed in the CNN improved the determination accuracy for all five lettuce growth indices due to the ability of the stereo camera to extract height information on lettuce. The average time for processing each lettuce image using the developed CNN model run on a Jetson SUB mini-PC with a Jetson Xavier NX was 0.83 s, indicating the potential for the model in fast real-time sensing of lettuce growth indices.

IGrow: A Smart Agriculture Solution to Autonomous Greenhouse Control

Agriculture is the foundation of human civilization. However, the rapid increase of the global population poses a challenge on this cornerstone by demanding more food. Modern autonomous greenhouses, equipped with sensors and actuators, provide a promising solution to the problem by empowering precise control for high-efficient food production. However, the optimal control of autonomous greenhouses is challenging, requiring decision-making based on high-dimensional sensory data, and the scaling of production is limited by the scarcity of labor capable of handling this task. With the advances of artificial intelligence (AI), the internet of things (IoT), and cloud computing technologies, we are hopeful to provide a solution to automate and smarten greenhouse control to address the above challenges. In this paper, we propose a smart agriculture solution named iGrow, for autonomous greenhouse control (AGC): (1) for the first time, we formulate the AGC problem as a Markov decision process (MDP) optimization problem; (2) we design a neural network-based simulator incorporated with the incremental mechanism to simulate the complete planting process of an autonomous greenhouse, which provides a testbed for the optimization of control strategies; (3) we propose a closed-loop bi-level optimization algorithm, which can dynamically re-optimize the greenhouse control strategy with newly observed data during real-world production. We not only conduct simulation experiments but also deploy iGrow in real scenarios, and experimental results demonstrate the effectiveness and superiority of iGrow in autonomous greenhouse simulation and optimal control. Particularly, compelling results from the tomato pilot project in real autonomous greenhouses show that our solution significantly increases crop yield (+10.15%) and net profit (+92.70%) with statistical significance compared to planting experts. Our solution opens up a new avenue for greenhouse production. The code is available at https://github.com/holmescao/iGrow.git.

Developing a multi-label tinyML machine learning model for an active and optimized greenhouse microclimate control from multivariate sensed data

In the uncertainties within which the worldwide food security lies nowadays, the agricultural industry is raising a crucial need for being equipped with the state-of-the-art technologies for a more efficient, climate-resilient and sustainable production. The traditional production methods have to be revisited, and opportunities should be given for the innovative solutions henceforth brought by big data analytics, cloud computing and internet of things (IoT). In this context, we develop an optimized tinyML-oriented model for an active machine learning-based greenhouse microclimate management to be integrated in an on-field microcontroller. We design an experimental strawberry greenhouse from which we collect multivariate climate data through installed sensors. The obtained values' combinations are labeled according to a five-action multi-label control strategy, then used to prepare a machine learning-ready dataset. The dataset is used to train and five-fold cross-validate 90 Multi-Layer Perceptrons (MLPs) with varied hyperparameters to select the most performant –yet optimized– model instance for the addressed task. Our multi-label control approach enables designing highly scalable models with reduced computational complexity, comprising only n control neurons instead of (1 + ∑nk=1Cnk) neurons (usually generated from a classic single-label approach from n input variables). Our final selected model incorporates 2 hidden layers with 7 and 8 neurons respectively and 151 parameters; it scored a mean accuracy of 97% during the cross-validation phase, then 96% on our supplementary test set. The model enables an intelligent and autonomous greenhouse management with the less required computations. It can be efficiently deployed in microcontrollers within real world operating conditions.

Parallel Control of Greenhouse Climate With a Transferable Prediction Model

Highly intelligent greenhouse without human intervention is the goal of autonomous greenhouse control. In this paper, a parallel control framework for greenhouse climate is proposed which aims to minimize the need for monitored data and expert knowledge. GreenLight climate model is used as a knowledge-based model that produces simulated data. LSTM with control units is pre-trained with these data. Test on necessary data size is done by transferring the model to other greenhouses. The new transferred model has a good improvement in the prediction of indoor temperature, humidity and CO2 concentration with approximate 0.05, 0.05 and 0.1 of R2, respectively, which shows the feasibility of the transferable prediction model.

Optimal smart contract for autonomous greenhouse environment based on IoT blockchain network in agriculture

The Internet of Things (IoT) has been widely adopted in many smart applications such as smart cities, healthcare, smart farms, industry etc. In recent few years, the greenhouse industry has earned significant consideration from the agriculture community due to its ability to produce fresh agricultural products with immense growth and production rate. However, labour and energy consumption costs increase the production cost of the greenhouse by 40–50% approximately. Moreover, the security and authenticity of agriculture data, particularly for yield monitoring and analysis, is also a challenging issue in current greenhouse systems.The greenhouse require optimal parameter settings with controlled environment to produce increase food production. Therefore, slight advancement can bring remarkable improvements concerning the increase in production with reduced overall cost. In this work, we contributed blockchain enabled optimization approach for greenhouse system. The proposed approach works in three steps to provide optimal greenhouse environment that are; prediction, optimization, and finally controlling. Initially, the Kalman filter algorithm is employed for predicting the greenhouse sensor data. In next step, the optimal parameters are computed for the indoor greenhouse environment. Finally, the optimized parameters are utilized by the control module to operate and regulate the actuator’s state to meet the desired settings in the indoor environment. To evaluate the performance of our proposed greenhouse system, we have developed an emulation tool. The proposed system has been investigated and compared against baseline approach concerning production rate and energy consumption. The obtained results reveal that the proposed optimization approach has improved the energy consumption by 19% against the prediction based approach and 41% against the Baseline scheme. Furthermore, the proof-of-concept based on the Hyperledger Fabric network is implemented on the top of the proposed greenhouse platform. For experimental analysis, we have conducted a series of experiments using Hyperledger calliper concerning throughput, latency, and resource utilization. These results advocates the efficiency of the proposed optimal greenhouse system.

A Simulator-based Planning Framework for Optimizing Autonomous Greenhouse Control Strategy

The rapidly growing global population presents challenges and demands for efficient production of healthy fresh food. Autonomous greenhouse equipped with standard sensors and actuators (such as heating and lighting) which enables control of indoor climate for crop production, contributes to producing higher yields. However, it requires skilled and expensive labor, as well as a large amount of energy. An autonomous greenhouse control strategy, powered by AI algorithms by optimizing the yields and resource use simultaneously, offers an ideal solution to the dilemma. In this paper, we propose a two-stage planning framework to automatically optimize greenhouse control setpoints given specific outside weather conditions. Firstly, we take advantage of cumulative planting data and horticulture knowledge to build a multi-modular simulator using neural networks, to simulate climate change and crop growth in the greenhouse. Secondly, two AI algorithms (reinforcement learning and heuristic algorithm) as planning methods are applied to obtain optimal control strategies based on the simulator. We evaluate our framework on a cherry-tomato planting dataset and demonstrate that the simulator is able to simulate greenhouse planting processes with high accuracy and fast speed. Moreover, the control strategies produced by the AI algorithms all obtain superhuman performance, in particular, significantly outperform all teams of the second “Autonomous Greenhouse Challenge” in terms of net profits.

Cherry Tomato Production in Intelligent Greenhouses-Sensors and AI for Control of Climate, Irrigation, Crop Yield, and Quality.

Greenhouses and indoor farming systems play an important role in providing fresh and nutritious food for the growing global population. Farms are becoming larger and greenhouse growers need to make complex decisions to maximize production and minimize resource use while meeting market requirements. However, highly skilled labor is increasingly lacking in the greenhouse sector. Moreover, extreme events such as the COVID-19 pandemic, can make farms temporarily less accessible. This highlights the need for more autonomous and remote-control strategies for greenhouse production. This paper describes and analyzes the results of the second “Autonomous Greenhouse Challenge”. In this challenge, an experiment was conducted in six high-tech greenhouse compartments during a period of six months of cherry tomato growing. The primary goal of the greenhouse operation was to maximize net profit, by controlling the greenhouse climate and crop with AI techniques. Five international teams with backgrounds in AI and horticulture were challenged in a competition to operate their own compartment remotely. They developed intelligent algorithms and use sensor data to determine climate setpoints and crop management strategy. All AI supported teams outperformed a human-operated greenhouse that served as reference. From the results obtained by the teams and from the analysis of the different climate-crop strategies, it was possible to detect challenges and opportunities for the future implementation of remote-control systems in greenhouse production.

Remote control of greenhouse cucumber production with artificial intelligence – results from the first international autonomous greenhouse challenge

The world population is increasing rapidly together with the demand for healthy fresh food. Greenhouse industry can play an important role, but encounters difficulties finding skilled staff to manage crop production. Artificial intelligence (AI) reaches breakthroughs in several areas, however, not yet in horticulture. An international competition on “autonomous greenhouses” aims in combining horticultural expertise with AI to make breakthroughs in fresh food production with fewer resources. Five international teams with different background in horticulture and AI participated in a growing experiment. Each team had a 96-m2 modern greenhouse compartment at WUR to grow remotely a cucumber crop ('Hi-Power') during a 4-months period. Each compartment was equipped with standard actuators and sensors, heating, ventilation, screening, artificial lighting, fogging, CO2 supply and irrigation. Climate control set points were remotely determined by teams using own AI algorithms, actuators were operated by a climate and irrigation process computer. Additionally, teams sent instructions for the crop pruning strategy. Measurements and set points were exchanged via a digital interface. Achievements in AI-controlled compartments were compared with a reference compartment, operated manually by three commercial growers. Teams used additional individually chosen sensors such as RGB or thermal cameras, temperature-humidity-light-sensor-networks, root-zone sensors, sap-flow meters or crop-weighing sensors. Teams' strategies for remote control ranged from supervised, unsupervised and reinforcement machine learning. Teams were judged and received points on several aspects: 50% for net profit, 20% for sustainability indicators, energy- and water-use-efficiency and CO2 consumption, 30% for their artificial intelligence algorithms, novelty, level of autonomous control, robustness and scalability. The results obtained by different teams in terms of climate and growing strategies, different resource use efficiency and net profit are presented in this paper. One of the AI-controlled compartments achieved overall better results than the manual grown reference compartment.

Evolutionary layout design synthesis of an autonomous greenhouse using product-related dependencies

AbstractThe development of autonomous greenhouses has caught the interest of many researchers and industrial considering their potential of offering an optimal environment for the growth of high-quality crops with minimum resources. Since an autonomous greenhouse is a mechatronic system, the consideration of its subsystem (e.g. heating systems) and component (e.g. actuators and sensors) interactions early in the design phase can ease the product development process. Indeed, this consideration could shorten the design process, reduce the number of redesign loops, and improve the performance of the overall mechatronic system. In the case of a greenhouse, it would lead to a higher quality of the crops and a better management of resources. In this work, the layout design of a general autonomous greenhouse is translated into an optimization problem statement while considering product-related dependencies. Then, a genetic algorithm is used to carry out the optimization of the layout design. The methodology is applied to the design of a fully autonomous greenhouse (45 cm × 30 cm × 30 cm) for the growth of plants in space. Although some objectives are conflictual, the developed algorithm proposes a compromise to obtain a near-optimal feasible layout design. The algorithm was also able to optimize the volume of components (occupied space) while considering the energy consumption and the overall mass. Their respective summed values are 2844.32 cm3, which represents 7% of the total volume, 5.86 W, and 655.8 g.

Autonomous greenhouse microclimate through hydroponic design and refurbished thermal energy by phase change material

The rapidly increasing demand for agricultural food needs coupled with rising energy costs marked challenges for ensuring food bigger than any previous period. Therefore, greenhouses agriculture is a main focus for farmers, engineers and greenhouses designers in view of ways of improvement that provides. Compared to the conventional greenhouses, hydroponics warrant better quality, greater nutrient content, higher yield, efficient water and fertilizer use, but also more energy requirements. A new Hydroponic Greenhouse was designed and installed in the Research and Technology Center of Energy in Tunisia. A Solar Air Heater with Latent thermal storage using Phase Change Material was also realized for the hydroponic greenhouse heating taking into account the thermal heat loads of the specific application. In this work, the microclimate of the Hydroponic Greenhouse without heating was pursued to evaluate the performance of the hydroponic design. Hydroponic greenhouse allowed better environment than conventional greenhouses. During daytime, the new greenhouse temperature exceeded 18 °C and the difference between the inside and the outside reached mostly 6 °C. The relative humidity ranged between 20 and 35% the day and 70–85% at night.Several measurements were also carried out after the heating to pursue the Solar Air Heater contribution. The temperature of the hydroponic greenhouse during nighttime after the heating raised by 6 °C and the nocturnal temperature was mainly over 15 °C. The diurnal temperature of the Heated Hydroponic Greenhouse was generally higher than 32 °C. Compared to conventional solar heating, the two packed beds of latent storage energy improved the indoor greenhouse environment especially during harsh and nocturnal periods.

Autonomous Greenhouse Sprayer Navigation Using Automatic Tracking Algorithm

Abstract. This article presents an accurate driving algorithm for autonomous greenhouse sprayer navigation. Due to the geological complexity of ground inside the greenhouse, the sprayer requires a stable driving algorithm under non-uniformal road conditions. We designed a body frame with an optimal size for the narrow aisle in a greenhouse. The driving algorithm was carried by an automatic turning algorithm under no-rail circumstance to maximize the utilization of a greenhouse space. Also, in order to segregate the turning algorithm and the straight path navigation, a photo sensor was additionally installed on the vehicle to detect the beginning of a turning area. The driving algorithm consisted of two kinds of autonomous procedures, which are the guidance rail tracking and the turning without any rail. The experimental results proved a driving function with an optimal turning radius, and a stable tracking performance on an unpaved clay road. It confirmed an increased slip rate of the vehicle’s wheel on the off-road situations, under the influence of a reduced rail width and an increased driving speed. Despite the severe driving conditions, the autonomous sprayer’s driving motion proved a stable tracking performance.

Deployment of a Fully-Automated Green Fluorescent Protein Imaging System in a High Arctic Autonomous Greenhouse

Higher plants are an integral part of strategies for sustained human presence in space. Space-based greenhouses have the potential to provide closed-loop recycling of oxygen, water and food. Plant monitoring systems with the capacity to remotely observe the condition of crops in real-time within these systems would permit operators to take immediate action to ensure optimum system yield and reliability. One such plant health monitoring technique involves the use of reporter genes driving fluorescent proteins as biological sensors of plant stress. In 2006 an initial prototype green fluorescent protein imager system was deployed at the Arthur Clarke Mars Greenhouse located in the Canadian High Arctic. This prototype demonstrated the advantageous of this biosensor technology and underscored the challenges in collecting and managing telemetric data from exigent environments. We present here the design and deployment of a second prototype imaging system deployed within and connected to the infrastructure of the Arthur Clarke Mars Greenhouse. This is the first imager to run autonomously for one year in the un-crewed greenhouse with command and control conducted through the greenhouse satellite control system. Images were saved locally in high resolution and sent telemetrically in low resolution. Imager hardware is described, including the custom designed LED growth light and fluorescent excitation light boards, filters, data acquisition and control system, and basic sensing and environmental control. Several critical lessons learned related to the hardware of small plant growth payloads are also elaborated.

Use of solar energy for the agricultural greenhouses autonomous conditioning

Capillary polypropylene exchangers are used to attenuate the differences between the diurnal and nocturnal air temperatures under the tunnel greenhouses. The climate conditioning system tested is composed of exchangers buried into a greenhouse and others suspended in its internal air. Water circulates in these exchangers in a hydraulic closed circuit. During the day, the suspended exchangers recover the energy in excess for the plant comfort. This recovered energy is stored into the greenhouse ground through the buried exchangers. During the night, thermal energy already stored is brought back by the suspended exchangers to heat the air in the greenhouse. A theoretical approach was elaborated to size thermal energy storage and restore system. The necessary exchange surfaces for the greenhouse conditioning depend on the required values of diurnal and nocturnal temperatures. The autonomous greenhouse conditioning was possible only for low temperatures. Temperatures of 20°C in the day and 12°C at night were reached in winter. More favorable temperatures for the plant growth of 25°C during the day and 18°C at night were reached during the other months. This conditioning system allowed a heat quantity extraction of 150 W.m −2/d. This energy quantity is sufficient to provide the needs of the tunnel greenhouse with polyethylene covers that is very common in Tunisia. The system cost is appropriate for agricultural applications. In addition, a surface quality and a constituting matter of this exchanger prevent the tartar formation and therefore its obstruction.