Design and assessment of an automated sweet pepper seed sorting machine

Abstract

Each year, large volumes of sweet pepper seeds (Capsicum annuum L.) are exported from Thailand to many countries. As a result, seed quality is very important. Sweet pepper seeds are very small, both in dimension and mass, therefore they are very difficult to inspect manually. Thus, we designed and constructed an automated sweet pepper seed sorting machine. It consists of three main stations: 1) seed feeder, 2) seed inspector using a camera and 3) seed sorter using a vacuum suction system. To transport seeds through each station, we designed a circular rotating plate with radial rows of blind holes. This allows all three stations to process a row of seeds in parallel. To identify discolored seeds, images were captured and processed to detect dark areas on the surface of the seeds. Our machine accurately sorted 90.9% of input seeds. The maximum potential capacity (in terms of seed processing rate) was 16,500 seeds per hour and we achieved actual rates of approximately 12,700 seeds per hour or roughly twice as fast as an experienced operator. Thus, this machine is faster and more accurate than human operators and can replace them.