Study on citrus fruit image data separability by segmentation methods

This paper presents the study on half of the citrus fruit image data acquired from a cold mirror bi-camera acquisition system. The ultimate objective of this project is to extend a ripeness study on the citrus fruit image data and the segmentation methods by multispectral analysis for fruit picking robot. To acquire the combination of the image data, an economic portable cold mirror acquisition system has been prototyped to align two CCD cameras with a classical cold mirror on a custom built fixture. The use of the cold mirror system is an attempt to capture different waveband images without algorithmic registration on images from the same channel by triggering and synchronizing two cameras closely. With interchangeability, some physical optical filters can be interchanged on the cameras as well to capture the attenuated image data. In the study, some segmentation methods have been selected conceptually covering the color indices, Fisher linear discriminate analysis, and the hyperplane construction using single perceptron with the multilayer perceptron, and the competitive self-organizing map. Therefore both the original image data and the attenuated data by physical optical filter from one camera are compared to discuss the data improvement and performance of the methods on segmentation. From the results, the polarizer filtered data is closely better than neutral density filtered data and much better than the original image data. Performance of the segmentation methods and some issues regarding the data are discussed in this paper.