Wheat Image Correction for Feature Segmentation Based on Color Linear CCD

Abstract

This paper presented a machine vision system with a color linear CCD to identify several China wheat classes. However, these images were deformed due to color linear CCD itself, image noise and uneven brightness also existed. The image needs to be corrected so as to segment the wheat features better. Algorithms of linearity equation were developed to correct geometric distortions, homomorphic filtering and space domain enhancement methods for image restoration also played a significant role. From the evaluation, we could find that the image treatment was likely to effective.