Gilching, Germany—Worldwide, the mushroom industry is a booming business valued at more than $50 billion in 2019, with most production originating in China and the United States. To create a more favorable environment for mushroom cultivation, farmers are moving their outdoor farming operations into large greenhouses, shifting environmental factors from external climates to internal microclimates. A major driver has been global warming, which has made outdoor farming less productive.
Controlling a greenhouse climate for optimal mushroom growth presents a number of challenges for farmers. A successful harvest requires frequent temperature changes, for instance. The greenhouse temperature must be reduced from +22°C to +16°C to stimulate fruiting. Moreover, relative humidity has to be maintained between 85% and 90%, and carbon dioxide at each of the six growth stages must be adjusted to appropriate levels.
For the most part, farmers must draw on their personal experience and visual observations to estimate the relationship between mushroom growth and the greenhouse microclimate and accordingly determine temperature, humidity, and other environmental factors. Greenhouses are equipped with environmental monitoring systems to obtain microclimate data. However, there is no sensor that can directly measure the growth of mushrooms to determine when adjustments should be made.
Recently, computer vision has been applied to agricultural technology (AgTech), most notably, deep-learning neural networks that can recognize individual images among multiple objects, such as a single mushroom cap in a large field of mushrooms. Harnessing the power of neural networks, scientists at Meiho University in Pingtung, Taiwan have developed a machine vision algorithm that can analyze images of the entire fruiting period of individual mushrooms. By continuously recording the size of individual mushroom caps, this data can be used for analysis to optimize greenhouse microclimate control, to calculate growth rates, and to act as harvest reminders.
During the experiment, an SVS-Vistek eco445CVGE67 1.2-MPixel color GigE Vision camera with image resolution set at 1296 × 964 pixels and 30 frames-per-second captured mushroom images. Besides accurate imaging, the camera was selected because it is rugged enough for a greenhouse climate which presents similar stresses as outdoor installation, including humidity and heat. The eco445CVGE67 is IP67 rated, so it is dust-tight and waterproof, as well as resistant to vibration because of its M12 connector.
Scientists installed cameras 25 cm above a mushroom cultivation bed with a light source placed at the top of the greenhouse. The cameras automatically began measuring the growth of the mushrooms during the fruiting period, calculating circle diameters of the mushroom caps every hour. Data was multiplied by spatial resolution to obtain the actual size of the mushroom circles.
Researchers conducted experiments and then compared their results with traditional imaging methods to validate the applicability of their proposed algorithm. They concluded that the new algorithm can successfully analyze images of the entire fruiting period of the mushrooms, making it extremely practical to farmers since it can make is easier to control the microclimate. They plan to expand their experiments to test the algorithm in large commercial applications.
For more information visit www.svs-vistek.com.
- "A novel image measurement algorithm for common mushroom caps based on convolutional neural network," Chuan-Pin Lu, Computers and Electronics in Agriculture, April 2020
