Smart greenhouses using internet of things: case study on tomatoes

In plants, photosynthesis occurs mainly within the leaves. Since photosynthesis requires carbon dioxide, water, and sunlight, all of these substances must be obtained by or transported to the leaves. Carbon dioxide is obtained through tiny pores in plant leaves called stomata. Oxygen is also released through the stomata [8]. Water and nutrition are obtained by the plant through the roots and delivered to the leaves through vascular plant tissue systems. Sunlight is absorbed by chlorophyll, a green pigment located in plant cell structures called chloroplasts. 6CO2+6H2O+ Chlorophyll+light→C6H12O6+6O2+ 6H2O(Carbon dioxide) (Water) (leaves) (Glucose) (Oxygen) \matrix{{6{\rm{C}}{{\rm{O}}_2} + 6{{\rm{H}}_2}{\rm{O}} + \,{\rm{Chlorophyll}} + {\rm{light}} \to {{\rm{C}}_6}{{\rm{H}}_{12}}{{\rm{O}}_6} + 6{{\rm{O}}_2} + \,6{{\rm{H}}_2}{\rm{O}}} \hfill\cr{\left( {{\rm{Carbon}}\,{\rm{dioxide}}} \right)\,\left( {{\rm{Water}}} \right)\,\left( {{\rm{leaves}}} \right)\,\left( {{\rm{Glucose}}} \right)\,\left( {{\rm{Oxygen}}} \right)} \hfill\cr}

If we took this equation as basic to growth in the chosen plants, then we must study each input of this equation and determine what the variables are that affect each input.

Ventilation is one of the most important components of a successful greenhouse. If there is no proper ventilation, the [9, 10] greenhouse and its growing plants can become exposed to severe problems. The main purposes of ventilation are to regulate temperature and humidity to an optimum level, ensuring air movement and thus preventing the buildup of plant pathogens (e.g., Botrytis cinerea) that favor still air conditions. Aeration also ensures that fresh air is provided for photosynthesis and plant respiration, and may enable important pollinators to gain access to the greenhouse crop.

The volume of water needed by tomato plants for optimal growth and production varies with the season and size of the plants. Soil moisture can determine how much water should be added, and soil moisture sensors can measure the appropriate amount of water needed. Young seedlings usually need about two fluid ounces (50 ml) per plant per day, while mature plants need two to three quarts (2 L) per day. On average, 90% of the water taken up by [11] the plant is used for transpiration (cooling), and the remaining 10% is used for growth. Inadequate amounts of water lead to wilted plants, and wilted plants are not productive. Prolonged water deficits can lead to permanent wilting, which can result in the death of the growing tips of the plant.

The water should be provided in six to twelve (or more) applications per day, depending on the type of growth medium used. Well-drained media, such as rice hulls, may need more than twelve applications per day. The timing of [12] applications should be determined by the water demand of the plants, with most of the applications occurring during the day. Some nighttime watering may be needed during periods of low humidity. More applications may be needed on sunny days to accommodate the increased rates of transpiration by the plants. Enough water should be applied that some drainage occurs from the growth medium bags (10%–20% of the water applied), and this additional amount of water needs to be added to the total amount of water required.

Light is an essential factor in maintaining plants. The rate of growth and the length of time the plant remains active depends on the amount of light it receives. Light energy is used in photosynthesis, which is the basic metabolic process of a plant. When determining the effect of light on plant growth, there are three areas to consider: intensity, duration, and quality. Light conditions and duration are not included in this research study, as we will make our test in optimum light duration and exclude light variables from this study. The concentration will be on plant nutrition [13, 14].

Increasing the time (duration) that plants are exposed to light can be used to compensate for lower light intensity, as long as the plant's flowering cycle is not sensitive to day length. The increased light duration allows the plant to produce enough food to survive and grow. However, plants require a period of darkness to develop properly and must be exposed to light for no more than 16 hours a day. Too much light is as harmful as too little. When a plant gets too much direct light, the leaves become pallid, sometimes burn, turn brownish, and then die. Therefore, we need to protect the plants from excessive exposure to direct sunlight during the summer months [15, 16, 17].

The geographic location plays a central role in managing the variables mentioned above [42, 43]. Temperature and humidity in our testing location, for example, are very close to what we needed inside the greenhouse for some seasons of the year. The length of the light (day, night) is mainly responsible for photosynthesis.

Tomato fruits have several stages of growth, starting from seed until collecting ripe fruits. Figure 3 defines the growing stages.

Figure 3

This research project is based on data reading, supervising, and control of tomato plant growth. In our experiment we will plant one tomato plant in a container. In Figure 4 the dimensions are defined using an inch unit (“). This container will be filled with soil. The whole container filled with soil will be placed on a balance scale and the weight calibrated to zero. Tomato seed will be planted in the soil to start the process. Each day in the morning the weight sensor will send data on the weight of the growing plant: only the plant, with all its contents, such as seeds, leaves, stems, flowers, and fruits.

Figure 4

The testing area of this research is in Aroub, which located in a suburb of Hebron, a city in Palestine (see Figure 5 for location).

Figure 5

This is considered a rich agricultural area that produces all kinds of fruits, especially tomatoes and cucumber. The annual climate condition in this area is given is given in Figure 6.

Figure 6

As we declared in the introduction to this article [44, 45], there are two parts of supervising and controlling greenhouses. Firstly, the surrounding conditions, which can be managed using given data of sensors on the basis of events – that is, the trigger method. Secondly, management of data concerning nutrition: more precisely, NPK added substances. The surrounding conditions aspect is more simple and straightforward. After testing location climate, we will start planting at 20 April, as this time is ideal for planting tomatoes in that location, as all variable conditions (light, humidity, temperature) are ideal for planting. We will do the planting process in the best light duration for planting tomatoes: at least 8 hours and 9 hours for the fruit formation and mature fruit stages, and no more than 9 hours in vegetative and flowering stages. Daytime ideal humidity levels in a green house range from highs of 90% down to 80%.

The nighttime moisture levels, however, range slightly lower, between 65% and 75%. The least humidity is needed at the fruit formation and mature fruit stages, and the most humidity is required during vegetative and flowering stages. The best temperatures for tomatoes at various stages is as follows: in-earth growing (12–17°C), vegetative (14–17°C), flowering (16–20°C), fruit formation (18–22°C), and mature fruiting (26–30°C). In any case, temperatures should not go above 28°C or below 5°C. The plant will die if temperatures soar above 30°C or drop to less than 0°C. In any case, if the RTH curve hits the upper or lower threshold, it will be a trigger to correct the issue, as mentioned above in the section on surrounding conditions.

Adding nutrition is very important for the plant in all stages mentioned above, from the nursery stage until mature fruiting stage. Basically the nutrients that we will consider in this research are nitrogen, phosphorate, and potassium (NPK), normally, [46, 47, 48], as we have discussed above. We will add to our container the following concentration values to reach better production: N (100–250 mg/kg), P (60–140 mg/kg), and K (60–120 mg/kg). We can control the concentration of these nutrition elements using soil sensors.

The soil nutrient content can be measured using an NPK soil sensor and Arduino of ammonium content. It is necessary to determine how much additional nutrient content should be added to the soil to increase crop production. The soil fertility can be measured using the NPK sensor. The major elements of soil fertilizers are nitrogen, phosphorus, and potassium. By reading values of soil nutrient concentration, we can learn about the nutritional deficiency or sufficiency in soil for plant growing stages. There are many methods for measuring soil nutrient elements, such as optical sensors or spectrometers. But the spectral analytical method of determining the nutrients is only 60% to 70% accurate. While comparing the spectral analysis method with traditional wet chemistry methods, accuracy of the product is yet to be fully resolved. So here we will use an optical sensor to detect nitrogen, phosphorus and potassium in the soil, as shown in Figure 7.

Figure 7

This kind of sensor is low-cost, quick, responsive, high precision, and portable, and it works with network and data transitions module Rs 485, with an output signal which can wirelessly transmit data up to 1200 meters [49]. The advantage of the optical sensor over a traditional detection method is that it gives instant measurements and highly accurate data. All we need is to insert its probes in soil and get the reading using Arduino. Thus, we can easily make our own instrument soil check NPK meter. In our test case we can read the details about the interfacing of soil nutrient sensor or NPK sensor with Arduino and display the NPK data in an OLED display. This soil sensor is suitable for detecting the contents of nitrogen, phosphorus, and potassium in the soil. It helps in determining the fertility of the soil and the sensor can be submerged in soil for a long time. It has a high-quality probe rust resistance, electrolytic resistance, and salt and alkali cover resistance. Therefore, this sensor is most appropriate for all kinds of soil. The sensor has high measurement accuracy, fast response speed, and good interchangeability, and it can be used with any microcontroller [50, 51, 52]. The sensor operates on 9 to 24 volts, and power consumption is very low. The sensor's accuracy is within 2%. The measuring range is from (0–1999) mg/kg. The operating temperature is between 5 to 40 °C. The output signal can be read using a mode for Rs 465 transmitter. The sensor has a predictive layer made using IP Success rating [53]. The sensor we need to make a NPK meter has four wires and works using mode unit 20 Amber converter along with twelve-volt supply.

The system that is shown in Figure 8 has two objectives. The first is to get an instant reading of nutrition concentration in units of mg/kg. This reading will be displayed on an OLED display. The second is to send data, transmitted wirelessly, to a server in a data center, for advanced control and data management to get the best NPK percentages [54, 55] and the most accurate percentage of NPK needed, depending on the growing phases defined in Figure 3. The main objective of big data collected is to control NPK added amounts to increase production of tomato fruits, for more than 90 days, which represents the life cycle. Each day we need to take the measurements of NPK concentration in soil and the weight of the whole plant. We can change the amount of NPK nutrition and see the resulting production, so we will then find the best conditions (values for amount of variables N, P, and K). Then the interface module needs to be connected to the NPK sensor to get the NPK reading. Then we assembled the circuit of the entire system, which can be applied on a breadboard. Apart from the breadboard, we will need any Arduino board, which can be programmable. Similarly, we used an OLED display square where we can display the NPK readings.

Figure 8

This system works to retrieve the data from the RS 485 mode bus from the NPK sensor, which has 3 lines: one line is for getting the nitrogen value in soil. [56] Similarly, the second line is for phosphorus, and the third is for getting the value for potassium. The instruction frame of data consists of the following data sets: –

Address code: the address code of instruction frame

The general form of the instruction frame is:

An example of response to inquiry in the case of nitrogen is:

We can calculate the soil nitrogen from the

Response received. For example, if we get 0030 as a response then soil nitrogen value will be: 0020 H(hexadecimal)=32 (Decimal)=> Nitrogen = 32mg/kg 0020\,{\rm{H}}\left( {{\rm{hexadecimal}}} \right) = 32\,\left( {{\rm{Decimal}}} \right) => \,{\rm{Nitrogen}}\, = \,32{\rm{mg}}/{\rm{kg}}

The NPK Sensor has 3 different inquiry frames for reading the value of nitrogen (N), phosphorous (P), and potassium (K). The inquiry frame is provided laterally with the instruction manual. For the NPK data the following individual inquiry frameworks:

Nitrogen: {0x01,0x03, 0x00, 0x1e, 0x00, 0x01, 0xe4, 0x0c}

Now, we are ready to read data and transmit the data frame to the big data repository. Using Arduino we can read data instantly from OLED and save it into the data center for future management and control.

We have seen in the previous section how to install the control system hooked with the NPK sensor. Programmable Arduino is connected for enabling an instant read of NPK amounts on a daily basis. Two main jobs are needed from the Arduino program: the first is to display the instant reading of NPK concentration in the soil; the second is to send signals to the cloud data center of all readings of given variables [57, 58]. Data transmission can be programmed on daily or hourly (or any specified) time intervals; in our case the data transmission can be on daily basis and no less. A per-day interval is chosen because, as we discussed before, real growth in plants normally occurs at night, which comes as an output of the photosynthesis process.

The interfacing software of the soil NPK sensor is connected to Arduino. It can retrieve soil nutrient values from the sensor via instruction command. The instruction command can be sent, and the value obtained in hexadecimal code. Then the HEX code needs to be converted into number values in decimal to get the value of soil nutrient contents in shape of big data [59, 60, 61].

Here we are using the OLED display to see the soil nutrient values (nitrogen, phosphorus, and potassium) in mg/kg. Also, we need to download the OLED Library. And added to this, the Arduino IDE commands can instruct a set of devices to do the following:

Change the value in one of its registers on/off, which is written to one bit register coil and read the value of holding register (16 bit) in Arduino

A Modbus instruction command contains the Modbus address of the device it is intended for (1 to 247). The Modbus instruction address is also called an inquiry frame. Only the addressed device will respond and act on the command, even though other devices might receive it, and it is transmitted to the big data repository. We applied the NPK data management on two experiments with ideal surrounding conditions, as mentioned in previous sections (geographic location at Aroub, Figure 5; start planting at April 25; and using the container mentioned in Figure 4 with same weight and soil; but in the first case it added NPK using soil mixing the traditional way of planting, which mixes soil with NPK; the second case in the experiment is to irrigate a solvent of NPK as a solution on the plants in containers, and check the amount of each nutrient on a daily basis. The experiment ran for 100 days [63].

Figure 9 shows two experimental plantings of tomatoes in two different containers in the same greenhouse conditions. In the first experiment we planted tomatoes using a traditional planting method: this method adds nutrients according to best-practice manual experience, which is adding nutrition as needed. The result – production of green and fruits – is up to 5.6 kg. The second part of Figure 9 shows planting tomatoes using controlled addition of nutrition based on each stage of growth. We found that with IoT controlled NPK addition, production will reach 7.4 kg, which exceeds the first, traditional, method by 2 kg. We can do more readings and experiments, playing with all variables, keeping records in our big data repository, until we define the very optimum production of greenhouse tomato cultivation.

Figure 9