A new algorithm based on hue saturation intensity（ HSI） space and learning vector quantization（ LVQ）neural network is proposed to improve the recognition accuracy of male tassel. The first image in maize field by RGB color space converted to HSI color space and extract its H component diagram

and the H component image is processed by binarization processing median filtering and connected region area denoising

morphological processing and generation of rectangular box connected domain

the threshold segmentation algorithm has been able to roughly identify the male tassel target. This paper intends to add validation steps in recognition algorithm

extraction of color and morphological characteristics are carried out on the identified candidate male tassel and the above characteristics is input into the trained LVQ neural network to confirm

the obtained segmentation results can effectively improve the accuracy of identification of male tassel

the accuracy reaches 96. 9 %.