Spot spraying is a new method. It is based on a concept that only the infected areas of the crop should be sprayed and not the whole field. This novelty can be implemented by having a modern sprayer, electronics and the target to spray in the crop. This last requirement is provided by the image annotation. Image annotation is the process of marking the areas of interest on images in order to make the computer able to recognize them in real conditions. Therefore, this combination leads to precision spraying with various benefits in crop production, human health and environmental sustainability.
Instead of spraying the whole field, spot spraying doesn’t harm the crop, but only kills the infestations. The same goes to the harmful weeds among the arable crops. Also, new sensors have allowed sprayers to treat each plant uniquely. These sensors identify plants directly and create an immediate treatment response. Image annotation is the basis. Regarding the economic aspects, spot spraying contributes to suppress the costs of the chemicals, because of the lower consumption during the spraying. Moreover, there is an important advantage in human health as most of the field remains clear without being sprayed. It is widely known that the vast majority of the farmers do preventive spraying as a measure to avoid the emergence of an infestation or the spreadance from the infected areas of the whole field.
The environmental benefits of the spot spraying stem from the lower chemical consumption on the one hand and the specific sprayed areas on the other. In regards to the first, lower chemicals reduce the probability of soil and water contamination. As far as the latter is concerned, it is true that traditional spraying may cause spray drift, potentially leading to rivers and lakes and in this way entering the water cycle. There is also the matter of high spray pressure which happens in the traditional spraying, while in spot spraying this is absent.
Image annotation is at the cutting edge and simultaneously the basis of computer vision. It is performed by the humans who have the domain expertise to do this task. Also, many AI-assisted annotation models have been developed as a faster way to annotate hundred or even thousand images. There is also an innovative technique, known as Active learning and is quite promising. Summarizing all the above we are able to understand the benefits to humans and agriculture and the strategy to solve many existing and emerging challenging tasks.