Nonlinear Fusion of Multispectral Citrus Fruit Image Data with Information Contents

Abstract

The main issue of vison-based automatic harvesting manipulators is the difficulty in the correct fruit identification in the images under natural lighting conditions. Mostly, the solution has been based on a linear combination of color components in the multispectral images. However, the results have not reached a satisfactory level. To overcome this issue, this paper proposes a robust nonlinear fusion method to augment the original color image with the synchronized near infrared image. The two images are fused with Daubechies wavelet transform (DWT) in a multiscale decomposition approach. With DWT, the background noises are reduced and the necessary image features are enhanced by fusing the color contrast of the color components and the homogeneity of the near infrared (NIR) component. The resulting fused color image is classified with a C-means algorithm for reconstruction. The performance of the proposed approach is evaluated with the statistical F measure in comparison to some existing methods using linear combinations of color components. The results show that the fusion of information in different spectral components has the advantage of enhancing the image quality, therefore improving the classification accuracy in citrus fruit identification in natural lighting conditions.

Remote Sensing Image Fusion Algorithm Based on Two-Stream Fusion Network and Residual Channel Attention Mechanism

“…(3) Spectral angle mapping (SAM) [51] reflects the degree of spectral distortion and calculates the angle between the corresponding pixels of the fused image and the reference image on the triad. SAM = 0 means that there is no spectral distortion:…”

Plant phenomics: an overview of image acquisition technologies and image data analysis algorithms

“…Traditionally, both multispectral and hyperspectral imaging have been used for remote sensing and have an increased number of applications in phenomics. A multispectral system has been developed to improve the original colour of images for fruit recognition [ 65 ]. The authors fused the original colour image with an infrared image using nonlinear Daubechies wavelet transform (DWT).…”

“…The features extracted from an image are disposed in the so-called “feature vectors.” The construction of feature vectors uses a wide set of methods to identify the objects in an image. The main features are edges, intensity of image pixels [ 39 ], geometries [ 158 ], textures [ 154 , 159 ], image transformations, e.g., Fourier [ 160 ] or Wavelet [ 65 , 161 ] or combinations of pixels of different colour spaces [ 131 ]. The end goal of feature extraction is to feed up a set of classifiers and machine learning algorithms (see below).…”

Desarrollo de un sistema de fenotipado basado en visión artificial para el estudio de la cinética de crecimiento de plantas

“…Tradicionalmente ambos sistemas se han empleado en el campo de la teledetección, y con la llegada y auge de la fenómica el número de aplicaciones se ha visto incrementado. En un ejemplo de aplicación, un sistema multiespectral fue empleado para mejorar las imágenes originales para el reconocimiento de frutos [97]. Aquí, los autores del trabajo fusionaron las imágenes en color originales con infrarrojo, utilizado transfomadas discretas wavelet de Daubechies (Daubechies DWT).…”

“…La construcción de un vector de características hace uso de una amplia variedad de métodos para identificar los objetos de una imagen. Las principales características son: bordes, valores de intensidad de píxeles [71], geometrías [205], texturas [201,206], transformaciones de la imagen como la transformada de Fourier [207], o la de Wavelet [97,208] o combinaciones de píxeles de diferentes espacios de color [122]. El objetivo final de la extracción de características es alimentar a un conjunto de clasificadores y/o algoritmos de aprendizaje automatizado (ML) (Sec.…”