Current limitations of Agriculture 4.0
Agriculture 4.0 has so far failed to live up to high expectations in terms of weed control. The key problem here is reliably distinguishing weeds from crops. The AI methods used for this purpose for image recognition are very complex, error-prone and only allow the recognition of one type of crop at a time - based on hundreds of thousands of images in different growth stages, weather and lighting situations. This high degree of specialization means that the first robots were designed exclusively for use on cotton, soybean, corn or fodder beet plantations in industrial agriculture. Nonetheless, such systems are unsuitable for vegetable cultivation or for smaller organic farms designed for multi-crop farming. However, it is precisely here that the manpower requirements are particularly high, making the use of robots especially promising.
Robotics in weed control is not possible without a cost-effective solution for universal and error-resistant differentiation between weeds and crops..
This is where biogriculture positions itself with a novel, patent-pending method of autonomous, mechanical weed control.