Using remote sensing technology to detect pests and diseases, Maifei Technology wants to be an AI service provider in the agricultural field
Gong Huaze, founder of Maifei Technology
my country's agriculture has a market size of over 100 billion yuan, but there are many pain points, such as pest and disease control. At present, the effective utilization rate of artificial plant protection pesticides is only 30%. On the one hand, due to insufficient personnel and lack of technology in plant protection stations, the forecast data is not accurate enough. On the other hand, farmers still use traditional manual and small sprayers to apply pesticides, and the degree of mechanization is very low. lower. Excessive pesticide application increases planting costs and can easily cause crop safety hazards and even lead to yield reductions.
Currently, for the prevention and control of pests and diseases in the market, satellite monitoring is available, and the signal is relatively stable, but the timeliness and resolution are insufficient, and it is easily affected by the atmosphere. Although drone plant protection can respond more quickly to farmland pesticide application needs, the accuracy of monitoring is generally low. .
Facing the big blue ocean of agriculture, Gong Huaze founded Maifei Technology at the end of 2016. What the company does isUtilize core remote sensing technology and self-developed monitoring machines to collect and determine the degree of pests and diseases in fields, and use big data and AI systems to generate more accurate pesticide application prescriptions to complete drone plant protection. The current business is mainly aimed at B-side agricultural cooperatives and large farmland owners.
Use remote sensing technology to detect pests and diseases
Gong Huaze once conducted research on remote sensing technology at the Chinese Academy of Sciences for many years. The so-called remote sensing technology means that it can conduct non-contact remote detection, including detecting invisible light to the naked eye, and obtain more dimensional data. One of the major application areas of remote sensing is agriculture, but it is mostly used for macro-level monitoring.
In order to seek further implementation of remote sensing technology and explore how it can directly serve agriculture, Gong Huaze decided to start a business. However, the planting industry in rural areas is relatively primitive, and farmers’ acceptance of new technologies is limited. How to break through? Helping farmers with pest and disease monitoring and operations has become an entry point.
In addition to detecting the spatial distribution of pests and diseases, McFly Technology's remote sensing monitoring can also classify the degree of pests and diseases into five levels. The specific implementation method is to comprehensively apply various wavebands to detect spectral characteristics, and then monitor the characteristics of pests and diseases. For example, the wavelength is 650-750nm, which belongs to the red edge band and is very sensitive to pests and diseases. The level of pests and diseases can be judged from the color of the spectrum.
Spectral Pest and Disease Detection-Level 2 Diseases
Liu Long, CTO of Maifei Technology, told Chuangye.com that the team previously sampled a 1,000-acre farm in Hubei for field verification. The monitored disease levels were completely consistent with the actual situation. The average detection accuracy of infected crops reached 97%, and the average pesticide saving was It can be used above 50% and has accumulated more than 160 million sets of spectral data.
Big data + AI provides a one-stop solution for scientific pesticide application
At present, Maifei Technology's products are mainly at the hardware level, including sensors, monitoring machines and plant protection machines. In the future, they will be equipped with software integrating pest and disease algorithms. For drone plant protection, accurately obtaining the distribution and extent of pests and diseases and "prescribing the right medicine" are the company's core competitiveness.
Maifei Technology has established a big data model to improve the accuracy of monitoring and the effectiveness of pesticide application predictions through data annotation, combined with genetic algorithms and deep learning algorithms. Since crops will be affected by climate zones and rice growing areas, the team will further combine field detection to revise the model.
In order to find a comprehensive solution, Gong Huaze and his team spent most of last year practicing the implementation model. The first areas they chose to launch their business were Hubei, Jiangsu, and Liaoning provinces, and an operation center in Jingzhou, Hubei has been built.
The business model is that McFly screens orders first. There are two dimensions of concern: one is whether the type of crops and sowing period are consistent, and the degree of contiguous farmland. If it is more concentrated, monitoring and pesticide application costs can be saved; the other is the control of customer quality.
Then, field information is collected through monitoring equipment to obtain the distribution map and severity level of pests and diseases; at the same time, McFly's own agricultural technicians will detect the standard dosage of pesticides on the spot, provide high-quality real data to the AI system, and generate targeted data for different conditions. The pesticide application prescription map for the level of pest and disease areas; finally the operator completes the pesticide application.
Maifei AI planting system
In this way, Maifei Technology provides farmers with a one-stop pesticide application solution, helping farmers reduce planting costs. As of December 2017, Maifei Technology’s pesticide application services have provided a total of 1 million acres.
This model is relatively heavy, but in Gong Huaze’s view, it defines the company more as an agricultural technology “service provider”. The penetration rate of new technologies in agriculture is relatively low, but only by ensuring the final effect through overall solutions can we build brand awareness and accelerate market expansion.
Author: Zhong Xiaoyu007. Pay attention to cutting-edge technology and consumption upgrades. WeChat ID: zhongsy_520. When applying as a friend, please indicate your company position and purpose of visit.
xiaobang 