{"id":6720,"date":"2023-11-26T08:31:16","date_gmt":"2023-11-26T08:31:16","guid":{"rendered":"https:\/\/innovationisrael.org.il\/en\/?post_type=article&p=6720"},"modified":"2023-11-26T08:31:18","modified_gmt":"2023-11-26T08:31:18","slug":"an-agronomist-in-your-smartphone","status":"publish","type":"article","link":"https:\/\/innovationisrael.org.il\/en\/article\/an-agronomist-in-your-smartphone\/","title":{"rendered":"An Agronomist in Your Smartphone"},"content":{"rendered":"\n
Human driven agriculture has undergone major upheavals since people began to domesticate plants and animals to produce food, and the pace of changes has only increased with the technological development of the past decades. The 1950s and 1960s witnessed the “green revolution” where the first understandings of the science behind genetics led to the invention of ways to design different species of plants that would be more resistant to diseases and the weather. Each agricultural revolution increased mankind’s ability to produce increasingly more quantities of food for the earth’s population \u2013 a critical ability considering the constant population growth. Farmers find themselves repeatedly devoting their lives to finding ways to produce more food with less land, effort, and resources. The growing demand for food is accompanied by a decline in the land available to grow it.<\/p>\n\n\n\n
Professionals believe that we are in the early stages of a new agricultural revolution: precision agriculture. As part of precision agriculture, farmers are adopting today’s advanced technologies \u2013 Artificial Intelligence and machine learning \u2013 to contend with the growing need for food, with dwindling farming land, and with the challenges posed by climate changes.<\/p>\n\n\n\n
In response to the growing need for new solutions to existing problems, two new Israeli companies are seeking to harness these technologies in their endeavor to gain a deeper understanding of the reality in existing farming areas.<\/p>\n\n\n\n
To appreciate the revolution offered by Israeli technologies, it is necessary to understand how traditional farming worked in terms of real-time monitoring of events in the field \u2013 from planting and sowing to picking and harvest. <\/p>\n\n\n\n
To understand the state of the land and the crops they grow, farmers have always been compelled to rely on their eyes and intuitions gained from frustrating years of contending with weather conditions, pests, diseases etc. In other words, to know what was happening in the field, they needed to go out to their fields, check the plants themselves, and make decisions based on prior knowledge, estimations, gut feelings and, naturally, guesses.<\/p>\n\n\n\n
If, for example, a farmer saw an infected crop, he had to identify the problem, estimate how many plants had been affected, and decide on a course of action \u2013 how to treat the problem, whether to spray pesticide on one plot or the entire field, or maybe even uproot the plants to prevent the disease spreading. Each insight and decision had far-reaching ramifications on both the farmer himself and on the people in need of his produce.<\/p>\n\n\n\n
According to various estimates, the world’s farmers lose between 30%-40% of their crop every growing season. The scope of these losses is estimated at approximately 500 billion dollars a year. It is this reality, whereby all the world’s agricultural produce is dependent on knowledge and possibly flawed human judgement that agri-tech companies seek to change while harnessing the most innovative technologies that mankind has to offer.<\/p>\n\n\n\n
The Israeli company Taranis <\/strong>is one of these companies. Gershom Kutliroff<\/strong>, the company’s CTO, explains that the company’s vision is to transform the entire agricultural industry into a field that is based on genuine data i.e., based on an accepted understanding about what is happening and not based on intuition.<\/p>\n\n\n\n The advantages are clear: farmers who make use of the technology to analyze the state of their fields and crops will have greater capacity to decide on the correct course of action. This will result in smarter, more accurate and more economical treatment of the crops \u2013 healthier for both the crops and the environment.<\/p>\n\n\n\n Take for example the use of chemicals to treat pests discovered in a certain plot in the field. The smart farmer will know exactly which plants are affected and need treatment, thereby saving not only the costs of using the pesticides, but also reducing harm to the other plants, and causing less damage to the environment, and increasing output.<\/p>\n\n\n\n Taranis was established in 2015 by Ofir Schlam<\/strong>, Eli Bukchin<\/strong>, Eyal Carmi<\/strong> and Assaf Horowitz<\/strong>. Kutliroff, who joined the company at a later stage, holds a doctorate in applied mathematics. He previously managed R&D teams in the field of computer vision and Artificial Intelligence and founded several companies.<\/p>\n\n\n\n Taranis, which operates in precision agriculture, offers AI-based solutions that help identify diseases and pests in a farmer’s fields. The company has developed a platform that uses Artificial Intelligence and machine learning to analyze very high-resolution images of fields received from UAVs, satellites, and light aircraft. The system’s algorithms are based on a unique bank of images compiled by the company. With over 200 million images, this bank is considered the largest in this area.<\/p>\n\n\n\n The farmers who use Taranis’s systems receive critical, real-time insights on the state of their crops: from identification of plots that are not growing properly because they need fertilizing and pest control to early identification of pests and diseases that are about to attack the plants. <\/p>\n\n\n\n Taranis, which recently raised over 100 million dollars, now employs 120 people in several countries. Its primary markets are the US and Brazil while the R&D is conducted in Israel.<\/p>\n\n\n\n “Once the season starts, we fly UAVs above farmers’ fields, gather images, upload them to the cloud and our servers, and then run AI-based models to identify and analyze the state of the crops”, Kutliroff explains.<\/p>\n\n\n\n Each field is photographed several times throughout the season. During the first flight, we photograph the number of plants beginning to grow to assess the need for re-sowing problematic areas in the field. Other tasks include searching for weeds, identifying the type of weed and enabling the farmer to only spray the affected areas with the right substances.<\/p>\n\n\n\n Subsequent flights search for more advanced threats in the field, including symptoms of diseases, a lack of potassium and nitrogen that can be identified by the color of the leaves or signs that something is unhealthy. All these flights are aimed at giving the farmer an in-depth understanding of what is happening in his field so that he can maintain it properly, thereby increasing the size of the harvest.<\/p>\n\n\n\n As part of its strategy, Taranis offers farmers a complete service as the company employs all the necessary pilots and drones. “In spring, for example”, Kutliroff explains, “there are more than 240 drones and 200 pilots. The company invests significant effort and capital in developing tools that enable maintenance of such a complex operation. Among other things, we have developed a special program to fly a UAV and ensure the high quality of the images etc. Each sophisticated UAV costs 30 thousand dollars but this enables us to supply the kind of service level and quality we expect”.<\/p>\n\n\n\n The images produced by the company are high-zoom and very high resolution \u2013 approximately 20 megapixels, with each pixel covering a third of a millimeter. “We can even see bugs on the plant!”, Kutliroff says.<\/p>\n\n\n\n Artificial Intelligence is employed when analyzing the images. As Kutliroff explains: “In the beginning, we used to send the images from the UAVs to human labelers in India and other locations. In theory, we could have continued doing this, but it wasn’t economical. It takes a long time and is very expensive, especially considering that during the 4 months of the season, we expect to receive between 7-8 million images”. <\/p>\n\n\n\n Furthermore, it is difficult to control the quality of the images’ precision when working with human labelers. The world of agriculture demands high-level expertise and only experienced agronomists can look at an image and identify a specific disease. For a normal unskilled person, this would be almost impossible. In contrast, Artificial Intelligence implements techniques that guarantee the quality of analysis for every image it looks at.<\/p>\n\n\n\n “We are presently completing the Innovation Authority’s program after receiving funding for three years”, Kutliroff says. “It has been a very successful experience and we enjoyed working with the Authority’s representatives who came to learn what we do.<\/p>\n\n\n\n “All the AI techniques we use are based on things that can be understood from academia and research, but significant gaps are revealed when they are applied to develop a product. It was very important to solve the challenging technical problems with the help of the Authority”.<\/p>\n\n\n\n