Foresight

What it does

Foresight is a technology-integrated, tree-mounted, fruit-measuring device that seeks to respond to the problem of repetitive, manually intensive, and often inaccurate collection of fruit data during the growing season through intentional design.

Your inspiration

Currently, a third of all food produced is wasted throughout the food production, distribution and consumption stages. Large-scale methods such as satellites are less effective when hail netting is used, increasingly common due to climate change. Timing for harvest is critical and dependent on many factors, including fruit size. Commonly, fruit measurement occurs manually with callipers, highly repetitive for users and subject to variation, as the size of fruit varies throughout the height of the tree. There needed to be a comprehensive, time-efficient and accurate means of data collection to best inform users about the state of their crop.

How it works

Foresight is attached to a tree trunk through an adaptable mounting system responsive to different orientations for orchard layout and tree shape. Operating below the hail netting level, Foresight houses a Zed2i camera, which uses stereoscopic vision, a much cheaper alternative to LIDAR, to collect visual imagery. These camera systems are already used in agricultural practices, and are plug-and-play, making for simpler setup and maintenance. Images are processed through Machine Learning that segments the fruit from its surroundings, and then measures the radius. Information from each unit links using WLAN to the orchard office. Foresight utilises night-time photography as a means of reducing environmental variation and capturing data in a controlled manner. Excessive glare affects the accuracy required for fruit measurement, and being able to photograph at night with calibrated flash lighting mitigates the effects of daytime glare on data interpretation.

Design process

The idea was born through observational work done in the orchard environment. A subsequent 247 ways to measure fruit were identified on post-it notes, and arranged by their feasibility for its context. A narrowed series of approaches were validated by a local Ag-tech Institution, and captured insights from most major apple and kiwifruit growers in the North Island of New Zealand. With a vast set of iterations, resulting in over a hundred prototypes, Foresight was critiqued and guided by stakeholders, tested each growing season, and thoroughly reviewed by designers working in industry to support the feasibility of the idea and the practicality of its implementation. Immersive research allowed highly practical in-field learning, trialling and gaining empathy for users. The resulting design concept is white to best attract bees, has an angled camera port to reduce water build-up, and features large, distinct buttons for easy instruction - all based on design insights gained while in the orchard. A scalable, affordable and practical solution that uses a minimal, durable aesthetic to house a highly-capable camera and its operating hardware in the resulting design. Supporting the concept is its packaging and succinct instructional booklet.

How it is different

Foresight is a novel solution to fruit measurement, combining trusted technologies simply and cost-effectively, placed onto the tree to capture real-time data that allows for comparison of growth across the season in visual and numerical modes. Comparison to previous seasons also bridges the gap between staff’s historical knowledge of the orchard, enabling new farmers to build on this understanding. Night-time photography reduces lighting variation. Foresight utilises a timer to capture an image and then turn off again, capturing daily data at each unit. The device has the option of being solar-powered or utilising a battery pack. Made from recycled HDPE for chemical and environmental durability, buttons and all componentry are protected through protective membranes. The band used to secure Foresight slots through the body of the device for best security and can be replaced with most standard straps, allowing self-repair and maintenance for growers.

Future plans

Operating below the hail netting level, Foresight data collection can also be used as a ground truthing strategy to larger-scale data capture such as satellite imagery. Having camera units in place on select trees can also allow for the future detection of pests, disease, as well as positive factors such as bees and pollination rates. Foresight's resulting data could be used to recommend how much thinning should be done, that is, the removal of some fruitless early in the season to make sure they have enough space to grow to a reasonable size. The mounting system can be developed to fit onto different farming equipment to scale its use cases.

Awards

Foresight has been named a Paris Design Award Winner, with some entries to other awards pending results. The resulting thesis on this project was awarded on the Deans' List of Exceptional Theses for the university.