Fig 9 - uploaded by Jose Luis Hernández Hernández
Content may be subject to copyright.
Source publication
The vision by computer has become a very important tool and powerful in the area of agriculture and agronomy for monitoring and automatic handling of the different agricultural processes. Digital processing of images is used to segment and classify leaves in the corn fields of the Mexican fields making use of color models. The techniques of segment...
Context in source publication
Context 1
... LBP operator is unaffected by any monotonic gray-scale transformation which preserves the pixel intensity order in a local neighborhood, as can be seen in Fig. 9. Initially for the LBP algorithm you should only work with one image channel, normally worked in gray scale or an LBP is calculated for each channel; a pixel is selected which will be the axis of the analysis, an additional order of comparison is determined, which can be any that the user requires, as long as, is kept constant in all ...
Citations
... With the rapid development of artificial intelligence technology, different techniques are based on artificial vision for digital image processing and the implementation of the image classification model. Reference [22] proposed methodology that consists of five stages, as shown in Figure 1, i.e., image acquisition, preprocessing, segmentation, feature extraction, and classification, to find the damages caused by the cogollero worm in corn fields. stages, as shown in Figure 1, i.e., image acquisition, preprocessing, segmentation, feature extraction, and classification, to find the damages caused by the cogollero worm in corn fields. ...
Plant diseases such as drought stress and pest diseases significantly impact crops’ growth and yield levels. By detecting the surface characteristics of plant leaves, we can judge the growth state of plants and whether diseases occur. Traditional manual detection methods are limited by the professional knowledge and practical experience of operators. In recent years, a detection method based on deep learning has been applied to improve detection accuracy and reduce detection time. In this paper, we propose a disease detection method using a convolutional neural network (CNN) with multi-scale feature fusion for maize leaf disease detection. Based on the one-stage plant disease network YoLov5s, the coordinate attention (CA) attention module is added, along with a key feature weight to enhance the effective information of the feature map, and the spatial pyramid pooling (SSP) module is modified by data augmentation to reduce the loss of feature information. Three experiments are conducted under complex conditions such as overlapping occlusion, sparse distribution of detection targets, and similar textures and backgrounds of disease areas. The experimental results show that the average accuracy of the MFF-CNN is higher than that of currently used methods such as YoLov5s, Faster RCNN, CenterNet, and DETR, and the detection time is also reduced. The proposed method provides a feasible solution not only for the diagnosis of maize leaf diseases, but also for the detection of other plant diseases.
This work presents preliminary results of a novel approach to automatically determine the sex (female/hermaphrodite) of papaya fruit using artificial vision and machine learning techniques. In agriculture, hermaphrodite papaya is the fruit with the highest demand among consumers, so it is preferred to plant and cultivate this type of papaya. Currently, sex determination (sexing) is done through laboratory tests, which are expensive and require a long wait for results. There is still no work that relies only on computational techniques to support farmers in this task, which would make it easier to perform. With this target in mind, we have developed a computer model that analyzes samples of female and hermaphrodite flowers at prehantesis (FaP), starting from nine weeks after planting the seeds. The features decided to use were texture and contour for different classifiers during the training and testing phases, such as K-nearest neighbors, support vector machines (MVS), decision trees (DT) and convolutional neural networks (CNN).
In this work we developed our own dataset. It contains 520 images of female/hermaphrodite FaP to start a moderately robust model training. Until now, the best obtained model accuracy is 80% applying the CNN technique using the incoming image matrix as a feature (contour). We consider this accuracy to be satisfactory for the time being, given the size of the dataset used. However, we believe that this accuracy can be improved by testing with a larger dataset and by including other features.
This work shows the development of an algorithm based on image processing techniques and aims to determine the quality of Manila mango for export purposes. Currently in Mexico, in most of the mango producing states, the analysis of this fruit is done manually and the quality of the mango is determined by considering the state of maturity, cleanliness of the mango´s skin and size. During this process, the fruit is granding and then classified; in these tasks, human errors can occur due to fatigue. The consequences of these errors translate into losses for farmers, since for one fruit detected to be of poor quality, the entire lot is rejected. Therefore, any attempt made to support the determination of the quality of this fruit in an automated way will be of great support, especially for small and medium-sized agricultural enterprises (SMEs). The methodology employed uses digital image processing techniques by analyzing color and texture, calculating the mean of the components and using statistical methods (histograms and co-occurrence matrices) in the regions of interest, in addition to applying the support vector machine (SVM) algorithm to classify Manila mango based on maturity and peel damage. The algorithm presented here details a process of obtaining the image to filter it and identify the edges of the mango, the representation and manipulation through the histogram is used to improve the image without affecting aspects that may be relevant in the image such as contours, textures and intensity. These characteristics will be used later to determine the quality of the fruit. The results shown so far are satisfactory, reaching 86% Accuracy.
