FarmSense: Bug-Proofing Crops with Sensors & Machine Learning

The Significant Economic Impact of Insect Pests on Agriculture

The annual damage inflicted on agricultural yields by insect pests, such as the Japanese beetle, surpasses $100 billion globally, as reported by the USDA’s Agricultural Research Service. Coupled with the spread of plant diseases – often facilitated by these arthropods – agricultural production suffers a 40% loss each year worldwide.

FarmSense: A Novel Approach to Pest Management

FarmSense, an agtech startup located in Riverside, California, is dedicated to tackling the challenge of insect pests. The company develops optical sensors and employs machine learning algorithms to identify and monitor insect activity in real time. This immediate data access is central to their approach.

Their system aims to enable the prompt application of pest control measures, including insecticides or biocontrol agents, through early detection. Traditional monitoring methods, utilizing mechanical traps, can delay crucial information by 10 to 14 days.

Eamonn Keogh, a FarmSense co-founder, explains, “Certain insects have adult lifespans of only a few days. By the time a problem is identified, it may already be well-established and significantly more difficult to manage.” He continues, “Real-time awareness allows for targeted interventions, reducing pesticide use, labor costs, and crop damage.”

The Technology Behind Real-Time Insect Detection

The functionality of FarmSense’s technology is complex. Their latest sensor, the FlightSensor, is currently undergoing testing in almond orchards in Southern California, supported by a Small Business Innovation Research grant.

The inspiration for the FlightSensor stems from Cold War-era espionage techniques. Keogh describes how Russian spies utilized lasers and glass panes to detect vibrations from conversations, translating them into intelligence.

“The core idea was to detect a flying insect by its unique signature,” Keogh stated. “Imagine a bug passing a laser beam – you would isolate the sound of the insect alone.”

The FlightSensor employs light curtains and shadows within a tunnel, attracting insects with specific attractants. A light source and an optical sensor are positioned on opposite sides. The sensor measures the amount of light blocked when an insect flies through the tunnel. This data is converted into audio and analyzed using machine learning algorithms in the cloud.

FarmSense emphasizes that the sensor is designed to resemble older analog devices for ease of use by growers and effectively filters out ambient noises like wind and rain.

“The signal quality is remarkably clear, and it’s unaffected by typical field sounds,” Keogh notes. “Listening to the audio from the sensor is akin to hearing a mosquito or bee in flight.”

Expertise and Cost Considerations

Eamonn Keogh, a professor of computer science and engineering at UC Riverside, specializes in data mining and develops the machine learning algorithms used by FarmSense. He is supported by entomologists and field specialists, including co-founder Leslie Hickle.

Shailendra Singh, the company’s CEO with a background in wireless and cellular networks and security systems, oversees the hardware development. He estimates the cost of each sensor at $300 per season.

The potential impact of this technology is substantial. Real-time insect data could significantly benefit farmers, enhancing their financial stability and promoting the conservation of vital resources like soil health.

FarmSense also aims to support small-scale rural farmers, who are disproportionately affected by insect-related crop losses.

However, the $300 per sensor per season cost could hinder adoption, potentially limiting the technology’s ability to address the broader issue of insect damage.

Risk Management for Small Farmers

Michael Carter, director of the USDA-funded Feed the Future Innovation Lab for Markets, Risk, and Resilience and a distinguished professor at UC Davis, highlights the challenges small-scale farmers face in managing risk.

“Risk can perpetuate poverty,” Carter explains. “It discourages investment in technologies that could increase income, due to the inherent uncertainty of the future.” He adds, “Farmers with limited resources cannot easily risk their savings on investments that might fail.”

Despite this, Carter is optimistic that technologies like the FlightSensor could reduce investment apprehension, particularly if coupled with insurance to provide further protection.

Beyond Real-Time Identification: Addressing Root Causes

The technology also prompts the question of whether real-time identification is the most effective pest management strategy. Research entomologist Andrew Liebhold of the USDA Forest Service suggests it may not be.

Liebhold emphasizes that the primary drivers of invasive insects – those causing the most significant damage to agriculture and forests – are travel and trade.

He supports the use of technology for controlling insect establishment but believes the optimal approach is to address the problem at its source. This includes reviewing current import and export regulations, pest treatment protocols, and potentially implementing travel restrictions.

Looking Ahead: Broader Applications and Future Potential

Despite these considerations, FarmSense’s technology holds considerable promise. Beyond addressing financial insecurity for farmers and safeguarding global food supplies, it could be valuable in tracking and disseminating information about disease-carrying insects, such as mosquitoes.

Given the ongoing disruptions caused by events like the COVID-19 pandemic, awareness of biosecurity’s importance – and the consequences of its failures – is heightened.

With projections indicating a 36% increase in non-native insect invasions by 2050 and growing population pressures on food production, innovative technologies like the FlightSensor, which enhance our ability to understand and respond to threats, are highly valuable.

As Carter concludes, “We need to be creative in exploring all potential benefits that agtech can offer to agriculture.”