CONCEPTUAL DESIGN OF AN INTELLIGENT COWPEA (Vigna unguiculata) HARVESTING MACHINE

Abstract

Harvesting of cowpea is one of the major operations in the cowpea production. It is carried out when the pods are fully matured and dried but these pods do not mature at the same time because of it staggered flowering period. Traditional handpicking of cowpea is still widely employed by most Nigerian cowpea farmers to harvest the rain fed cowpeas intermittently; while the late maturing cowpeas are left to dry completely and then harvested at once using cutlass or sickle. Manual harvesting continues to be one of the most time consuming and labour intensive, resulting in low efficiency and limited competitiveness. These however, pose significant obstacle to sustainable food security that only cutting-edge technologies such as agricultural robotics and computer vision can address. This work is aimed at Conceptual Design of an Intelligent Cowpea (Vigna unguiculata) Harvesting Machine using a Single Shot Detector (SSD) deep learning algorithm. In this work, multidisciplinary conceptual design methodology adopted, and consists of five different stages established towards intelligent cowpea harvesting machine. The first stage is determination of the engineering properties of the cowpea plant, pod panicles and dried pods relevant to the design of an intelligent cowpea harvester. This work, investigated and presented the overall framework and high-level ideas towards the development of an intelligent cowpea harvesting machine. The outcome of the study presented a feasible design concept as a prelude to solution of challenges confronting manual hand picking of non-uniformly maturing cowpea pods.