AN INTELLIGENT COMPUTER VISION BASED APPROACH TO MONITOR PHARMACEUTICAL PRODUCTION LINE

Urooj Arif ¹, Faizan Naseem ²  

 Abstract —— The idea of An Intelligent Computer Vision Based Approach to Monitor Pharmaceutical Production Line  is introduced by our Group after the events observed by us in different pharmaceutical companies cases studies. Like in Japan there they use laser scanners to scan the objects whether they are filled or empty if empty they are discarded and filled goes to next level of production. But in the economy of Pakistan Pharmaceutical industry is the most growing industry in Pakistan and a lot of companies are stepping in to this sector the basic elements to step in in this industry includes the idea, Research and Development , State of the art facilities for manufacturing and packaging etc. So we decided to work on the factor of manufacturing and packaging sector and to work majorly on production line to make production line intelligent by introducing the Computer Vision Based Intelligent Approach for the Production Line. And this product will be of high usability and applications in this industry in coming times specifically in Pakistan may be implemented to other Third world countries as well.

Keywords—-Image Processing, Artificial Neural Network

I. INTRODUCTION

The Project is about Object Detection Identification and Comparison using Computer Vision Based Embedded System. And as we know Object detection is the process of finding instances of real-world objects such as faces, Vehicles, and buildings in images or videos. Object detection algorithms typically use extracted features and learning algorithms to recognize instances of an object category. It is commonly used in applications such as image retrieval, security, surveillance, and automated quality control systems.

II. LITERATURE REVIEW

Pakistan is a very fast and future-oriented pharmaceutical industry. At the time of independence in 1947, was perhaps the pharmaceutical industry. Today Pakistan is about 400 pharmaceutical manufacturing units including those operated by 25 multinationals in the country. 70 per cent of the country’s pharmaceutical industry, meet the demand of finished medicine. The domestic pharmaceutical market is divided almost equally between citizens and multinational companies in exchange.
Progressive development of the pharmaceutical industry, especially in the last decade, as shown in years. The industry has invested heavily in recent years to upgrade, and Good Manufacturing Practices (GMP), the majority of the industry in accordance with national and international guidelines. Currently, the industry has the ability to produce a variety of products from simple to sophisticated biotech tablet, and value-added generic oncology compounds. Although Pakistan Pharma – health and without access to modern medicine for almost half the population, fastest growth. It needs a chance, but clearly much more work to be done by governments and economic stakeholders. Sales per capita consumption of less than US $ 10 worth of drugs every year $ 1.4 billion in 2007, and 2012 should be sold in 2012 represents 2.3 B drug prices,

Last year, the economy, due to poor economic conditions and suffered a lot of damage from the fight against terrorism. Moreover, the economy, inflation, slow economic growth and seen a significant reduction in the value of major currencies against the rupee. No doubt the Government will take the necessary steps to get out of this terrible situation, but that it is the economy and adaptation to board immediately rejected.
Before reviewing the pharmaceutical market, we look at the situation globally. 48%, 28% and the US, the EU and 12% of the value of the global market-based controls for Japan with a share. The remaining 20% of the global pharmaceutical market.The pharmaceutical market style, it is very competitive and challenging. Some 600 pharmaceutical companies in Pakistan. The 386 operating units. Have already mentioned, has a dramatic effect on inflation and devaluation operating margin. Therefore, the challenges for the pharmaceutical industry can be a major cause.When we look back in the pharmaceutical industry, the situation has clearly changed completely. Early nineties it was occupied by the multinational pharmaceutical companies in the industry but there is also a scenario for the past 18 years or so, I was completely changed. 30 386 units of multinational companies operating in drug production. Multinational and national companies in the ratio of 45% and 55%. Increasing share of national companies is an encouraging sign for domestic investors. The annual growth of 11%, US 1.64 $ billion total pharmaceutical market, pharmaceutical industry today with the global development of Pakistan. In the present scenario, more than 80 percent of demand met with national companies.
Local while 20 percent are covered by imports. All raw materials are imported from China, India, Europe, North America and other countries. About 20% of the total imports from Switzerland. 2008-09, according to GlaxoSmithKline (GSK) List Data List 11.59% market share and multinationals with growth of 8.9%. The national pharmaceutical companies Getz. (Pvt) Ltd has over 70%, with 3.76% market share in the last three years. And a 2.79% market share of the list of national pharmaceutical companies with Getz Pharma, Some multinational companies.

Currently, the government of Pakistan to take action drugs, drug manufacturing license of 4 species shall:
1. Creating
2. Basic manufacturing.
3. Half-based manufacturing
4. Repackaging

The organization of the census of Pakistan (2010), Pakistan has a population of 16.98 million with growth rate of 2.69. Everyday health consciousness also, as well, which can be helpful in increasing per capita expenditure growth in the future.The most successful pharmaceutical companies, existing business and customer trends to maximize profits with a variety of marketing strategies for different types of customers in almost constant new challenges and opportunities. Pharmaceutical companies invest (ROI), to improve their return New communication technologies (digital media) to its traditional ways.

III. IMAGE PROCESSING

Image processing can be set to produce either a picture or features or parameters that are related to image. Images are processed as a stereo signal, in which the third dimension or Z axis.
Image processing refers to a digital image processing. However, an analog and optical image processing is also possible. The pictures are said to achieve imaging. Image closely linked to computer vision and graphics processing. Born as images manually objects, physical models of the atmosphere and light in computer graphics, instead of some imaging devices having a camera on landscapes such as the purchase of the more animated movies . Computer vision contrast is often regarded as a high-level image processing from a machine, or a computer or software will decrypt the contents of the physical layout of the photo or image.
Because of the growing importance of technology and scientific concept of the modern science of images that are too wide. Examples include genetic research or real-time multi-asset portfolio microarray data into portfolio management in the financial sector.We also are different factors involved in image processing, can be from one to another can vary include, binary images, color photos (8 bit, 24 bit), Color Quantization, Convolution, Distance Matrics, Dithering, Edge Detectors, Frequency Domain, Gray scale Images, Image Editing Software, Idempotence, Isotrophic Operators, Kernels, Logical Operators, Masking, Non-linear Filtering, Pixels, Primary Colors, Spatial Domain, Structuring Elements, Wrapping and saturation.

IV. ARTIFICIAL NEURAL NETWORK

A neuron is a nerve cell is the basic building block of the nervous system. A large number of neurons in ways similar to other cells in the human body, but there is a significant difference between neurons and other cells. Neurons are specialized to transmit information throughout the body. This highly specialized nerve cells are responsible for the information in the chemical and electronic forms of communication. Different types of neurons, which are responsible for the different tasks within the body. Sensory neurons carry information from the sensory receptor cells in the body to the brain. Motor neurons transmit information from the brain-body muscle. Neuron body responsible for information and communication between different neurons.
Neuron matching neurons and other cells in the body with other cells, both of which are surrounded by a membrane that seals protect the genetic information of other cells of neurons and body are composed of a central can. The bodies of both types of cells in cell mitochondria, Golgi contains, including the organelles and cytoplasm body. Cell Life
Create unique to neurons, differences, unlike other body cells shortly after birth ceased to reproduce neurons. For this reason, some parts of the brain neurons will die because of the neurons at birth than life, but not the place. Neurons do not reproduce in most areas of the brain, have shown that the formation of new connections between neurons Research.Neurogenesis, or the formation of new nerve cells, about life is found in parts of the brain. Neuron is a membrane that is designed to send information to other cells. dendrites, cell bodies and axons are three main parts of a neuron. However, all neurons vary somewhat in size, shape and characteristics in terms of neuron function and role. Some neurons while others are highly branched, to obtain a large amount of information, there are few dendritic branches. Others may be quite a long while some neurons, short axons. The longest axons in the human body is about three feet in length and extends from the underside of the spine to the big toe! Neuron dendrites receive information from sensory receptors or other neurons. This information is then passed on to the cell body and axon. Once this is known as an action potential which travels down an electrical signal, the length of axons in the axon’s information.
Communication between synapses is an electrical pulse has reached the end of a Axon will be transferred once, information about the synaptic gap to the dendrites of neighboring neurons. In some cases, signals about the space between neurons and soon on its way. In other cases, are required to send information from one neuron to the next neurotransmitters. The neurotransmitters cross to reach other neurons and synaptic receptor sites vacuum chemical messengers that are released from the axon terminals. Neurons, neurotransmitters and their receptor binding site in a way known as reuptake are reabsorbed will be used again.

V. TRAINING AND TEST ING PHASE

We have break the our training and testing phase in to the steps that includes Input, Image Filtering, Image Segmentation, object detection, Visual feature, machine learning, trained network, classified output.

VI. METHODOLOGY

The basic Architecture is human tracking and monitoring are deployed on a visual application tool for testing, which serves as package for human classification and monitoring the trained proceeds and its validation is implemented in MATLAB to build a software system.

     i. Image Registration

Somebody entrance is the method of transforming to  incompatible sets of aggregation into one nascent system. Collection collects quadruplicate photographs,  information achieve from diverse sensors, present, depths, or viewpoints. individual enrolment old in computer modality method, medical images, biological images  military target recognition  the compiled and analyzed to images and data to collect from camera. Registration is important  to compare or integrate the data to achieve from these different measurements.

      ii. ROI Selection

Blobs are  supply to a extra statement of individual structures in spatiality of regions, the corners  more point-like. blobs are described to  preferred direction that capital  the more blob detectors are also be regarded in  touch part operators. Blob detectors can sensed areas of  human they to such glassy to be sensed in a carre four of device.

      iii. Edges

Edges is a line between two individual regions.  Line can be most of concretism, and countenance junctions. Edges are commonly is a sets of points in the mortal which soul a vehement  ratio. The  algorithms gift be constraint sharp circumstantial points together for   consummate description of  provide. These algorithms is  usually guess on the properties of an urgency,i.e mould, smoothness, and close continuance.

VII. EXPERIMENTAL RESULTS

An experimental setup was established to record the Objects with different angles we have followed the criteria of having a pic from 60 different angles and in this way we have worked out on 10 different tablets. In this way we have hypothesis of 600 images as our repository

VIII. CONCLUSION

After all the observations and research made on this project we reach to the conclusion that this project is an Intelligent  Computer Vision Based Approach To Monitor Pharmaceutical Production Line. This will be the most beneficial project for the pharmaceutical industry most specifically in Pakistan based economy and will get the benefits like  Smart remote Monitoring, Reliability, Increase Professionalism, Precise Detection, Approx. Zero Chances of Error, Detection  Counting and Data Backup All in one Time and most important no need of human workers in this system that makes this project different from the current standards of the systems that are still be followed in the pharmaceutical industries majorly in Pakistan.

VIIII. FUTURE WORK

We have achieved this project to this level as this report present but we are still looking to make this system more efficient by adding some more features or modules that we may be able to add with this hardware and can achieve more factors to be intelligent and automated.

REFERENCES

1. https://www.scribd.com/document/24495237/A-to-Z-of-Image-Processing-Concepts

2. https://www.scribd.com/document/325872322/Digital-Image-Processing-is-the-Use-of-Computer-Algorithms-to-Perform-Image-Processing-on-Digital-Images

3. https://www.scribd.com/document/266789152/Neural-Network

4. https://www.scribd.com/document/273031421/HYBRID-APPROACH-FOR-SEMANTIC-OBJECT-DETECTION-IN-VIDEO
5. https://www.scribd.com/document/161457953/SWOT-ANALYSIS

6. https://www.scribd.com/document/39266814/Industry-Analysis

7. https://www.scribd.com/document/63877699/Image-Processing