Search

Maintaining the purity of pig breeds is an essential task for their economic value. The traditional breed identification methods through coat colour are prone to error due to huge intra-breed variation. This paper uses principal component Analysis (PCA) to classify the pig breeds using their images. Individual images of five different pure breeds were captured from organized farms in India under both controlled and uncontrolled environments. Three different image sets were created, containing images in the controlled, uncontrolled, and mixed environment image sets. With 80:20 training to testing datasets, 93% accuracy was found in the proposed method of principal component analysis. Finally, two performance-based comparative analyses of our method were done with PCA-based methods and other renowned techniques used for animal breed identification, wherein our PCA method outperformed others in both comparative scenarios.

Hyperspectral Imaging has become an important method in monitoring vegetation and crops and producing information regarding onset of diseases. The availability of spectrometers in the market is not only costly but also requires standard-operating-procedures in using them. The use of Specim IQ, a handheld spectrometer with a wavelength-captivating-range of 400-1000 nm has been found to be optimistic in field-application with least complexity. In this paper, an early possible detection of downy mildew has been analysed through the difference in spectral distribution patterns through the reflectance-wavelength graph so that the well-advanced result may avert any significant damage over the visual eyes.

In this paper, an attempt has been made to develop a model to decide with precision the breed identity of individual goat by using its image. For image-based multi-class classification tasks, CNNs have been found to be the best tool. But selecting the most efficient CNN model for a particular classification scenario is a very difficult job. To find an optimal CNN model for goat breed prediction, we have compared two of the most popular pre-trained deep-learning-based CNN models (VGG-16 & Inception-v3) based on their performance. Both the models have been fine-tuned using transfer learning on the goat breed database. This goat breed database has been created from goat images of six different breeds, which have been captured from different organized registered goat farms in India and almost two thousand digital images of individual goat have been captured without imposing stress to animals. It has been observed that Inception-v3 has outperformed VGG-16 with higher accuracy and lower training time. To measure the prediction performance of this fine-tuned Inception-v3 model, it has been applied to a test set of pure breed goat images and standardized classification performance evaluation metrics have been used to evaluate the prediction results. From the results, it is established that the proposed method used in this paper is able to accurately classify (recognize) goat breeds with high accuracy. Finally, comparison has been made with prediction accuracies of different technologies used for identification of domestic animals.

Image classification using fully connected neural network is not efficient due to huge number of parameters in each layer. In this paper, we propose a Neuro-Statistic model for classification of five different pig breeds from pig images. The model consists of four sub modules which work together as a layered structure. We captured multiple individual pig images of five different pig breeds from different organized farms to conduct this research, segmented the captured pig images using hue based segmentation algorithm and then calculated the statistical properties like entropy, standard deviation, variance, mean, median, mode and color properties like H.S.V from the content of the individual segmented images. We fed all the extracted properties into Neural Network for Pig Breed (NNPB) to perform pig breed prediction with the classification module and analyzed the best performance, regression error plot, Error histogram and training state of NNPB. The performance of NNPB network was accepted based on error analysis and finally, we used the trained model to predict the breed of 50 pig images and achieved the prediction accuracy of 90%.

Accurate and reliable breed identification of domestic animals from images is one of the most promising but challenging tasks in intelligent livestock management. Traditional methods for animal breed identification are very costly and time consuming. Therefore, there is a need for a faster and cheaper technique for animal breed identification, which can be used by anyone without much technical knowhow. Deep Learning based animal breed classification from images can be used to solve this problem. Recent developments in deep Convolutional Neural Network (CNN) has drastically improved the accuracy of image recognition systems, but choosing the optimal model for the required task is very important for best performance. In this study, the performance of nine different deep CNN-based models have been analyzed to find the optimal model which can precisely determine the breed identity of individual animals from its image. All nine CNN models have been separately trained end-to-end on Pig Breed Dataset and Goat Breed Dataset using a set of identical hyperparameters. From the results obtained it has been established that MobileNetV2 is the best deep-CNN model for Goat Breed Classification with 95.00% prediction accuracy and InceptionV3 is the best model for pig breed classification with 100.00% prediction accuracy. Breed classification performance of goat and pig obtained in this study have been compared with other techniques used for animal breed classification. Comparison results show that our CNN-based technique has performed on par with all other methods. With these encouraging results, it can be confidently stated that deep CNN-based models can be used for solving the animal breed classification problem with high accuracy and can be used as ready to use technology for intelligent livestock management.

Agriculture is the primary source of livelihood to many countries around the globe and contributes to about 6.9% of the world’s total economic production and worth about $5,084,800 million. Therefore, progressive growth in agriculture is very much needed for these countries. The production rate of crops in agriculture is affected by a number of factors such as temperature, humidity, rainfall, the onset of pests, etc. which are not under the direct control of the farmers. The management of diseases, pests and the fertility of the soil can be proclaimed by the application of different types of pesticides, insecticides, fertilizers, etc. through manual spraying by the crop scouts over the hectares of land, which affects the nervous system and most of the functionality of the human body parts. In this chapter, the concept of implementing drones for spraying pesticides, fertilizers, etc., has been proposed for intelligent agricultural management. It assures the deduction of negative effect on the farmers and helps to stimulate the idea of handling fertilizers and pesticides in the areas. The drone in the form of a quadcopter has been designed and implemented with the technology of spraying pesticides for instantaneous action as soon as the disease onset is confirmed either manually or technically. The practical implementation of this technology would help in covering large areas of agricultural land in a small quantum of time and also reduce the cost of spraying widely. A wide application is the use of the drone in precision agricultural management where with pre-defined trajectory and measurements, the farm productivity and management can be improved with least manual labour and most optimistic results.

Animal identification is necessary for records, registration, and proof of ownership. The owner of few Black Bengal Goats can identify his goats by sight but it will create a problem for a larger number of goats as they are looking almost similar. A number of identification tools have been used for Black Bengal Goats like ear tag, tattoo, branding, RFID, etc. The Tattoos are permanent identification marking but inconvenient to read after a few months or years. Most of the farmers and breeders have used ear tags, which contain a number for identification of particular goat but may be lost at the time of grazing. Some organized farmers have placed RFID chips in tags but RFID reader is necessary to read the content of chips. In this paper, an effort has been made to identify individual Black Bengal Goat using their iris image like a human. The eye images have been captured preprocessed, enhanced, and irises have been segmented. The template has been generated from each segmented iris and stored in the database. The matching has been performed among different segmented iris images from the same goat and also been performed among iris images captured from different goats. It has been observed that the average Hamming distance among iris images captured at different times from the same goat are different from the average hamming distances among iris images from other goats. Finally, the matching threshold has been decided for the identification of Black Bengal Goat.

In this paper, convolutional neural network object detection technology has been used to detect pig breeds with high precision from images captured through mobile cameras. The pretrained model is retrained on several images of 6 different pure breed pigs obtained from organized farms. The Faster R-CNN Inception-ResNet-v2 model has been used in transfer learning fashion for the above task. The training accuracy of this model is 100%, and the testing accuracy of this model is 91% with a confidence level of 94%. From the results achieved, it is noted that this model has produced better results compared to detection accuracy on other datasets like dog dataset, flower dataset, etc.

In this paper, individual light coloured pig (Yorkshire) has been recognized based on their ear images captured using mobile phones. The ears have been kept parallel to the light source and the images have been captured from the opposite side of the light source. The auricular venation pattern from each captured ear image has been extracted, the template has been generated and stored in a database. The templates of recaptured ear images have been matched with the stored templates of the same pig using average Euclidean distance. The pig has been verified if average Euclidean distances of matching are \({

The study aims to focus on the entrepreneurship possibilities in the livestock sector (particularly Pig) in India. The existing individual pig identification system is tamperable, stressful, hurtful in nature resulting in low reliability among the insurance companies in providing the insurance claim as demanded by the farmers post animal’s death. Thus, a new method of identification has been developed, which is tamperproof, reliable, unique, and non-invasive. Due to a large number of Pigs in India (approx. 9.06 million), there is an entrepreneurship possibility to commercialize the identification technology in the market. The market has been studied and was concluded that charging Rs. 100 per biometric tag, revenue of about INR 90.6 Crores can be generated pan India. Finally, after analyzing the Lifetime Value and Cost of Customer Acquisition, the break-even point was found near year two and was predicted a huge profit on and after the second year.

Disease identification of the plant at an early stage is a key to prevent the major diseases by minimal application of chemical pesticides. In West Bengal, phoma blight is now becoming an emerging dreaded disease of potato. Phoma blight is associated with development of numerous spots on leaflets, thereby reducing green photosynthetic area causing huge loss in potato tuber production. The disease rating scale for severity analysis has not been developed for this disease till now. In this paper, an image-based phoma blight disease rating scale has been developed using k-means clustering. The image of phoma blight affecting potato leaflets has been captured using a DSLR camera by placing white paper background of leaflets. The percentage of affected areas has been calculated and an image-based phoma disease scale has been developed. The number of affected leaflet images has been given to several plant pathologists and they have assigned disease rating scores based on eye estimations. The score has been assigned for each leaflet considering the maximum number of same scores assigned by plant pathologists. The disease rating has also been assigned based on the actual affected area within each leaflet image using k-means clustering. The comparison has been performed between eye estimated scoring and k-means-based scoring to verify the scale. Finally, it has been observed that the new developed disease rating scale has given 87% accuracy in disease estimation with plant pathologists.

Convolutional Neural Network (CNN) is widely used as a strong framework for object classification. This framework is suitable for classification of group of objects which carry similar features like animal breeds, plant varieties, etc. In this paper, an attempt has been made to build a model of identifying individual Ghungroo pig (one of the most popular registered pig breeds in India) based on its face image using CNN. Individual Ghungroo pigs look very similar, and it is a very challenging task even for humans to identify them from a group of Ghungroo pigs. Ten Ghungroo pigs from ICAR-Indian Veterinary Research Institute (IVRI) Animal Farm Kalyani (Block B, Kanchrapara, West Bengal 741235) have been captured using a DSLR camera. The pigs have been captured in natural environment without any restriction. The Inception-v3 model with transfer learning has been used to build the individual pig identification model. The captured images have been separated with 90%:10% into training images and test images sets, respectively. It has been observed that the model has predicted individual pigs with 100% accuracy at 0.87 confidence level.