Farmica-Pak will primarily be used by the farmers and bioscience researchers. The system will be used by the farmers to detect and eradicate the weeds on their farms. Whereas the bioscience researchers will use the platform to enhance their information base concerning the weeds and obtain information in short interval.
Automated weed/crop discrimination is an active area of research where machine learning techniques can be applied that enable weed control strategies with weed specific treatment to save cost and mitigate environmental impact. We intend to develop an app that will be available on android and web platforms, which will automate weed/crop discrimination. Through this application we can get the details about weed plant by image classification and know the required amount of herbicide to use thus saving cost and the environment.
Agriculture is considered as the backbone of Pakistan's economy; it depends on crops. The most important crops are wheat, sugarcane, cotton, and rice. So, our purpose is to develop an app that can automate weed/crop discrimination. Through this application we can get the details about weed plant by image classification and know the required amount of herbicide to use thus it will save cost and save environment from pollution.
Traditionally agrochemical spraying often results in over and under dosing. Weed classification can help in variable rate spraying that will be cost effective and will give high per acre yield. We are creating a solution for bioscience researchers/farmers which is available “on click”. They will get all details about weed plant and their species by images classification and they will know which herbicide to use in the right amount by just a click on their mobile phone.
The developed solution is Artificial Intelligence powered and uses deep leaning-based Inception V3 neural network for the detection and classification of weeds. Additionally, the product contains the blogs and weed information section for the bio researchers so that they can have wide outreach towards the community. The developed solution is aimed at automating, streamlining, and making the sample collection process more effective and efficient.
