Automatic Recognition and Harvesting of Ripe Tomatoes under Greenhouse Conditions- A Survey

Abstract: There are numerous challenges in harvesting agricultural products and marketing them effectively. The challenges range from rainfall to availability of seasoned land to labour and marketing supply chain. In manual process, considerable time, hard labour and money is spent in inspecting and harvesting yield. This results in high cost of labour and additional time taken to complete entire process of harvesting, and holistically, deteriorates quality and consistency of the yield. The end-user market is quality demanding and expects consistency within one lot, and also among lots for them to be loyal to the brand and the produce. Often, this is not achievable for reasons discussed above. In many cases, where farm product is input to a final product, say tomato ketchup, high quality and consistency are essential to maintain the quality of the end-product, thereby guaranteeing good customer experience. This paper contains a viable engineering solution to this socially relevant problem of automated harvesting of ripe tomatoes under greenhouse conditions. Image-processing techniques, explained in this paper, are used to identify ripe tomatoes, and once identified; an automatic robotic vehicle harvests the ripe tomatoes.

Keywords: Image Processing, Watershed Algorithm, Robotic Harvester, Automated Harvesting, Greenhouse Navigation.