In order to meet the high efficient detection of the corn ear quality, a detection method of traits for corn ear were presented based on hue, saturation, value (HSV) color space. Firstly, the corn ear images with green background were acquired by using the machine vision technology, and then remove the green background using HSV histogram threshold algorithm, as well as filtrate sharp edges and noise using FFT filter. The particle filter was used to separate corns in an image. After four iteration corrosion by the corrosion morphology method, the 3 row between the ear of corn was obtained . The size, shape, texture and color characteristics were detected for corn ear. Using this method tested the 67 images of corn ear, the test results show that the testing accuracy of corn ear size and shape feature was 100%, while the ear color and the texture feature detection accuracy rate was 98.55% and 96.25%, respectively. The average detection time of one corn was 0.1 s.