mobile harvesting solutions

Smart Harvest consists of multiple hardware and software modules that together significantly improve the process of harvesting mushrooms in several ways.

Central to the system is the visual analysis module in which mushrooms are captured with a high-resolution camera on the mushroom bed. In these images, mushrooms are found, measured and the texture and colour are assessed. By comparing current and previous images, the growth of each mushroom can be modelled precisely.

This analysis makes it possible to provide quantitative and qualitative information on the mushrooms currently in the harvest cycle, but Smart Harvest goes further than that. A picking advice -based on a massive number of calculations and optionally adjusted
by an expert- can be projected directly on the bed. In addition, the entire harvest cycle can be planned and optimized by the system, making optimal use of available capacity and avoiding mushrooms getting picked too late or too early. All this results in a highly
optimized harvest yield.

Using proprietary algorithms developed by AMR, in which smart, self-learning techniques are used, it is possible to detect and model hundreds of mushrooms in a container within a few seconds. The colour and bulging of the mushrooms are part of this analysis
as well.

Picking advice

Using the values obtained by the visual analysis, the system then calculates dozens of secondary features for each mushroom which influence whether/when a certain mushroom should be picked. Think of the number of neighbours / contacts that a mushroom has, the number of mushrooms in a certain radius, or the average growth rate of the mushroom over the last few hours. A smart expert system can, based on this information, provide the optimal picking strategy at any given moment.

Using a projector, the picking advice is projected directly onto the bed, where mushrooms to pick are for example coloured red.
Mushrooms can also be sorted based on qualitative characteristics such as diameter, weight, shape or colour. The various types can be given a different colour during the projection.

The decision to pick or leave behind a mushroom at a certain moment greatly depends on when the next harvest opportunity is. The decision also influences the growth and optimal picking moment of many of the remaining mushrooms in the container.

Smart Harvest uses historical data to accurately predict the weight, diameter and quality of a mushroom at any point in the future and can anticipate the influence of neighbouring mushrooms getting picked.

Smart Harvest is able to model the entire harvest cycle, including the physical infrastructure, which means that the system knows exactly when each container will arrive at the picking station and may even be used to control the velocity of containers in the mobile harvest system and to sort containers in different lanes in order to influence the picking moment. With these two features and a complex planning and optimization algorithm, Smart Harvest maximizes the yield and product quality of the entire harvest.

Smart Harvest provides several ways of viewing the harvest results and controlling the system. On a touch screen / tablet, the main results can be seen and the main components can be controlled.

A desktop (Windows) application provides access to more extensive analysis results, such as comparisons of multiple harvests or different containers. Using the desktop application the user can inspect the results for an individual container or even an individual mushroom.

Finally, all main statistics can be accessed through a website from a remote location or on a mobile device.