Cattle Counting in an Image Test phase, srs, design phase and source code final deliverable
Domain/Category
Image Processing
Abstract / Introduction
Image processing refers to a set of techniques used to manipulate and extract valuable information from images, with the aim of improving their quality. This process is widely used in a variety of fields including medical science, defense, industry, remote sensing, pattern recognition, and video processing, to name a few. We can also use image processing for livestock management. Cattle counting is an important task for livestock management. Traditional methods of counting cattle, such as manual counting, are time-consuming and prone to errors. The objective of this project is to develop an image processing system that can count the number of cattle in an image with a high degree of accuracy.
This project proposal aims to develop an image processing system that can count cattle with high accuracy. The proposed project covers various aspects of image processing, including preprocessing, object detection and segmentation. The evaluation of the system on a dataset of images with varying levels of complexity will help to determine the accuracy and speed of the system. This project will contribute to the development of an efficient and reliable system for counting cattle.
Functional Requirements:
- Collect a dataset of images containing cattle in various settings, such as pastures and feedlots.
- The dataset should be consist of 200 different images with varying levels of complexity, such as images with varying lighting conditions, backgrounds and different numbers of animals.
- Preprocess the images to remove noise and distortions, and adjust the contrast and brightness.
- Develop an algorithm to detect cattle in an image using object detection techniques, such as the YOLO (You Only Look Once) algorithm.
- Apply segmentation techniques to segment the cattle from the background in the image.
- Develop an algorithm to count the number of cattle in the segmented image using various techniques such as counting the number of distinct cattle contours.
- Implement a machine learning algorithm (AlexNet CNN) to improve the accuracy of the cattle detection and counting.
- Develop an interface for users to input images and display the count of cattle.
- Randomly distribute the dataset in to 70% and 30%.
10.Train the system on 70% (140 out of 200) of the images.
11.Test the system on 30% (60 out of 200) of the images.
12.Evaluate the performance of the developed system, compare it to traditional methods, and determine the accuracy and speed of the system.
Note: Virtual University of Pakistan will not provide any kind of hardware for this project, student must arrange required hardware by himself/herself.
Tools & Technologies:
Preferred tool and technology: MATLAB (Any latest version of MATLAB)
Supervisor:
Name: Noor Rahman
