Why are current, common methods to estimate vegetable harvests in greenhouses prone to error and imprecise?
Current methods to estimate future harvests require time-consuming, manual measurements each week by a team of agronomists/Growers; measurements focus on counting plant characteristics, such as the number of under-ripened fruits and flowers. Coupling these plant characteristics with weather conditions, historical records and past experience, agronomists can achieve harvest estimates that are often over or under the actual harvest yield by 30% or more. Given that many environmental and biological factors influence plant growth, the simple measures used today for harvest forecasting are too limited.
Using artificial intelligence, Motorleaf can analyze a multitude of growth factors that determine when fruits and vegetables will ripen, and thus, achieve more precise harvest estimates. Moreover, our artificial intelligence technology of machine-learning algorithms actually learn and become more and more accurate over time when trained with additional data acquired from each harvest cycle.
Motorleaf’s technology enables greenhouses to automate harvest forecasts. On average, this automation will free up 1 to 3 days of work for agronomists that was required for manual harvest forecasts; they can now focus their efforts on business development and researching new plant varieties.
With precise harvest estimates, greenhouse owners gain new abilities to better manage labour by knowing beforehand how many pickers and packers they require to bring their harvests to market.
Marketing and sales professionals acquire better means to price produce and can better time promotions and discounts by knowing future supply and surpluses of produce.
Motorleaf strives to provide at least a 2x to 3x return on investment to customers purchasing our artificial intelligence technologies. There are many sources of value, cost savings and other monetary benefits from our harvest forecasts.
Here is a list of factors you should consider when evaluating the monetary benefits of automated, precise harvest forecasts:
- How much produce could you have bought at a better price from another farm if you knew weeks prior that you would be unable to produce enough of your own produce to meet contract obligations with your wholesaler?
- How much of your produce could you have sold at a higher price if you knew weeks in advance that you will have excess produce? This will avoid the need to sell excess, perishable produce as quickly as possible below market price.
- How much produce would you avoid wasting if you could ensure you had enough employees available to harvest and pack the fruit on time rather than it rotting onsite? How much money would you save from better labour management that avoids the need for employees to work overtime in order to harvest your crop on time?
- How much more money would you have made on a promotional discount for your produce if you could better match surpluses in production with the timing of the promotion?
- How many hours and employees do you devote each week towards calculating your harvest yields? What benefit would your greenhouse obtain if automated this task and used these employees to do other work in the greenhouse?
- What reputation and value do you gain by avoiding producing too little produce to meet contract obligations with a buyer? Will you be better positioned to renew future contracts with this buyer? If yes, would you be able to negotiate better prices for your produce if you are more than twice as reliable in your harvest forecasting compared to your competitors?
An algorithm is sophisticated software that conducts calculations, follows rules and/or solves specific problems. Algorithms are capable of processing large amounts of data and automating repetitive, logic-based decision-making.
Algorithms are the basic software components used in common categories of artificial intelligence, namely machine learning and deep learning. Most often, we will refer to Motorleaf algorithms as “automated greenhouse technology” or by the algorithm’s function, such as “automated harvest forecast calculator”.
Artificial intelligence is computer technology, being sophisticated software that can strengthen its ability to process information as the software learns from past and present situations. “Past and present situations” are information in the form of data. By presenting the software with large amounts of data–known as a “learning phase” or “training”–the software can find complex patterns within that data. Once the software determines the patterns, we can provide the software with new information and request that it predict a final result.
Using artificial intelligence, Motorleaf builds software that learns patterns in greenhouse data that can predict the growth of crops, and thus, when crops are ready for harvest. By training our software with your past greenhouse data, we can provide the software with your current growing conditions and have it make an accurate prediction of your future harvest yields.
To start developing your harvest yield service, you will need to share with Motorleaf your greenhouse data and historical records of past yields kept in spreadsheets or notebooks. We will then fine-tune your algorithm over time by importing additional grow data.
To develop the most accurate automated harvest yield forecasts, you need to provide Motorleaf with consistent historical data: your data must be from the same crop variety and the same climate zone in the greenhouse so we obtain information about the same type of plant growing under the same conditions. Submitting a mix of data for different crops, different growing conditions and different greenhouse zones will compromise the accuracy of the harvest forecasts.
No, Motorleaf does not sell or share your greenhouse data with anyone. As a subscriber to our automated harvest forecasts, Motorleaf and your greenhouse will sign a non-disclosure agreement specifying that we will not share any of your greenhouse data nor the quantities of your harvest yields.
In addition to our legally-binding mutual non-disclosure agreement, Motorleaf is also bound by its own internal governance regulations established in collaboration with our investors; our investors include two of Canada’s largest banks, Desjardins, The Business Development Bank of Canada, as well as the multinational corporations, DuPont Pioneer and Bayer (part of Radicle Growth).
Do you use my greenhouse data to develop products for greenhouse customers other than my own greenhouse?
No, Motorleaf does not share or use your data to develop products for anyone else. We do not merge (or ‘aggregate’) greenhouse data to develop our artificial intelligence technologies.
On average, Motorleaf requires 2 to 6 weeks to develop your harvest forecast algorithm. Development times will vary depending on the availability of greenhouse data and the quality of that data. We then conduct a testing period to prove you can automate harvest forecasts using our custom-made technology; this testing period requires tracking your harvest yields for at least one month.
At the very least, Motorleaf will sell our harvest forecast services only if they are accurate enough to enable your greenhouse to fully automate harvest forecasting. In addition to automating this task, Motorleaf’s harvest estimates are often 50%-70% more accurate (i.e., 50%-70% reduction in error in predicting harvest yields) than estimates made using current crop yield calculations.
Once Motorleaf develops harvest forecast services for a crop in my greenhouse, will that service predict the harvest yield for other crops in different areas of my greenhouse?
No, each harvest forecasting algorithm Motorleaf develops for your greenhouse will only be accurate for one specific crop variety in one defined area of your greenhouse. Each crop variety has unique attributes, and each greenhouse zone has distinct growing conditions that will influence the overall growth of that crop. Therefore, the greenhouse data needed to develop harvest forecasts is different for each crop variety and greenhouse zone.
In order to obtain equivalent harvest forecast services, you will need to provide Motorleaf with additional greenhouse data for each crop variety at distinct zones in your greenhouse.
Motorleaf charges a subscription for harvest forecasts, where we price our services based on particular attributes of your greenhouse (e.g., size of farm, quality of data available for developing the harvest forecasting algorithm, purchase of a yearly subscription versus a monthly subscription, request to develop our forecasts for several different crop varieties, etc.).
By continuously accumulating ever more greenhouse data from diverse data sources, Motorleaf can ensure your automated harvest forecasts remain accurate and become even more precise over time. For those reasons, your greenhouse may need to purchase additional sensors and data collection equipment. We recommend that all our greenhouse clients purchase the Motorleaf Space Sensor and Heart.
Our automation technology requires no robots or advanced robotic equipment.
Motorleaf takes great care towards protecting your greenhouse data. Your greenhouse data is processed onsite within a closed communication network; to protect the information within your data, we convert the real-time, raw data into 5-minute summaries; we calculate and store a summary of the most important criteria of the data, such as the maximum, minimum and average values within a 5 minute period. We then delete your original raw greenhouse data; only the summarized data is transmitted to our Canadian databases through Motorleaf’s secure API.
You maintain ownership of all your raw greenhouse data, while both Motorleaf and you own the summarized data. Motorleaf is ready to delete your summarized data on request; simply contact Motorleaf’s Customer Excellence department to make this request by phone 1(888)687-5301 ext: 702, or email: firstname.lastname@example.org.
Greenhouse customers typically upload their data to Motorleaf’s software platform, the Grow Journal. This process transfers data using a secure SSL connection, which ensures encryption of passwords and information, rendering it unreadable by 3rd parties across the internet. Both you and Motorleaf must remain careful to keep your password safe and secret. At any time, you may request changing your password and security information by contacting Motorleaf’s Customer Excellence department by phone 1(888) 687-5301 ext: 702, or email: email@example.com.