Thumbnail Image

Thesis

Browse

20201
Reset filters
Author: Pawanpreet Kaur × End date: 2020 × Start date: 2020 × Thesis Type: M.Tech. ×
Reset filters

Search Results

Now showing 1 - 1 of 1
  • Thumbnail Image
    ThesisItemRestricted
    Computer Vision to recognize maturity level of Peach cv. Punjab Nectarine
    (Punjab Agricultural University, Ludhiana, 2020) Pawanpreet Kaur; Derminder Singh
    Agriculture plays a key role in India's economic growth. Many new technologies need to be develop to place the agriculture sector at the frontline. Even now in India, human experts conduct the maturity inspection of fruits. It is really laborious and time-consuming process. In assessing agricultural produce, meeting quality standards and increasing market value, grading of fruits is necessary. Computer vision system and image processing techniques have been found to be highly useful in the food processing industry in recent years, particularly for quality check and sorting of agriculture produces. In the proposed work, digital image processing techniques are used to detect maturity level of Peach cv. Punjab Nectarine fruit using digital images of fruit taken at different stages of fruit development. Various image pre-processing techniques are applied on images. K-means clustering technique is used to divide datasets into 5 clusters and it performs color-based segmentation that is color based ROI. It finds out number of points in an image. After that, SURF algorithm is implemented to check what features are present in an image and then extracts that features from an image and store them in database in the form of key points. Its accuracy rate is more as compared to SIFT (Scale-Invariant Feature Transform) algorithm, takes less time for processing and extracts global data from an image. It extracts features like color, size, contrast, Mean, SD, Entropy, Homogeneity, Energy, and Correlation etc. using MATLAB (Matrix Laboratory) software. After that, the machinelearning algorithm MSVM is applied to classify image according to maturity level of fruit using various parameter values obtained. Results indicated that color and shape characteristics of fruit are important for visual inspection. The fruit size and color values calculated from desktop application are compared with manual readings to check accuracy of software. When the length and breadth computed from application were compared with the manual readings, the error percentage was less than 2.3% and in terms of L,a,b values the error percentage between manual and software results was less than 10.82%. Therefore, it is computed that software accurately detects maturity level of fruit with faster speed and performance. Method used in present software that is SURF+MSVM has high accuracy rate and fast computing speed. As per the past manual practice, maturity is calculated with a naked eye whereas this software will help in detecting maturity with quantitative value as compared to manual process.