The invention discloses a fruit and vegetable
visual detection system based on multispectral
light source imaging, and aims to solve the problems that an existing fruit and vegetable
visual detection system is limited by fruit and vegetable shapes and surface
color saturation changes of the existing fruit and vegetable
visual detection system, and high-brightness light spots are easily formed on the surfaces of
fruits and vegetables to be detected; and
image acquisition and subsequent analysis and detection results are influenced. The detection system comprises two
LED lamp sets which are arranged in a
bilateral symmetry mode and used for lighting the positions of
fruits and vegetables to be detected, and light reflecting plates used for refracting light to the positions of the
fruits and vegetables to be detected are arranged in the
light irradiation direction of the
LED lamp sets. A color camera and an
infrared camera which are used for performing surface imaging on the to-be-detected fruits and vegetables are arranged above the to-be-detected fruits and vegetables in parallel; and according to the detection system, visual detection is carried out on the surfaces of fruits and vegetables through multispectral lamp
source imaging, halo generated by direct light is avoided, and therefore the accuracy of information collection and subsequent analysis and detection results is improved, and high-precision visual detection is achieved.