Design of Day-Lily Robot Recognition System Based on ZYNQ

Abstract

With the development and wide application of robots in agricultural production, more and more agricultural fruit and vegetable robot schemes had been put forward, but the research on day-lily robot is less. Aiming at the problem of high labor cost in the current day-lily picking process, this paper designed a portable, low-cost embedded recognition system based on ZYNQ based on the characteristics of the ripe fruits of day-lily. A low-cost portable embedded recognition system was designed based on ZYNQ for day-lily robot. ZYNQ processing board was used to recognize the fruit image, and the combination of traditional image processing and machine learning recognition algorithm was used to realize the efficient real-time recognition of the target fruit. Traditional image processing combined with machine learning recognition algorithm could be transplanted to ZYNQ processing board which had the FPGA, which could better guarantee the real-time image recognition. Moreover, this recognition system has high recognition efficiency, suitable accuracy and low cost, which could meet the identification requirements of portable and low cost of agricultural robots at the same time.