Let’s resolve the biggest confusion between all comprehensive technical terms like artificial intelligence, machine learning and deep learning. Artificial intelligence is a broader space under which Machine Learning and Deep Learning are subsets of it. Let’s understand briefly about each of them to get a better picture. The concept of artificial intelligence came into existence in 1956. But data at that time was not sufficient to calculate accurate results. Artificial Intelligence Artificial Intelligence is a technique by which machines demonstrates intelligence or behavior like humans. In AI, machine can learn from experience, like new born kids do. So from new input data, machine can adjust new responses. We can consider artificial intelligence as a project of creating Huge Monument which can take centuries to build. So the one who started building it, could not even see it fully built. AI researchers started working on bricks and bases of the project by creating learning algorithms so that future researchers will use it to build smart intelligence system. Example of Artificial Intelligence: Apple Siri, Microsoft Cortana, Tesla Self Driving Cars and many more. Machine Learning In Artificial intelligence, it was difficult to train complex decision making operation models of the Human brain. “Machine Learning is an application Artificial intelligence which enables machine to learn from statistical data to improve with experience.” The designed algorithms in Machine Learning are developed in such a way, that it can learn and improve the results when new data is provided. Example of Machine Learning: Netflix, Google Maps Elaborate Examples: Netflix – Depending upon what type of movies and series you watch, Netflix will suggest you same type of movies and series to you in Recommended section. Google Maps – Google map analyses the traffic and suggests you the fastest routes to your destination. Deep Learning Deep learning is a part of a broader family of Machine Learning that is inspired by the functionality of our brain cells called artificial neural network. It takes data connection between all the artificial neurons and adjust it according to the data parent. With the increase in the size of data parent ,more neurons is added . You can relate Deep Learning as rocket a rocket engine which uses huge amount of data a fuel to process the algorithms. Deep Learning concept is not new but recently it’s hype has increased and getting a lot of attention. How Deep Learning Works at simple scale: In above example, machine will validate all the criteria to check if the rectangle is a square. When it is nothing but nested hierarchy of conditions and checks. Deep learning does the same thing but a larger scale Above blog gives the brief difference of difference between different types of AI. Developing an application requires in depth knowledge and understanding of AI. If you want any assistant with AI application development you can always contact us. Techaroha Team is specialized in Block Chain and AI Application. Techaroha is One of the Best Software company to build application with a mix of Block Chain and AI.
Why ?? Pothole are a biggest hurdle in the development of the country.They harm citizens in directly or indirectly in many ways.Pothole Detection System can solve the Problem faced by Every Citizens of our Country. 1. 10% percentage of Fuel is loss due to Potholes causing an unnecessary loss of 8.5 Billion that is 56 thousand cores 2. Death Due to Potholes in India – 3597 , which is much more compare to any other unnatural death cause in India 3. Millions of Man Hours are lost due to people stuck in traffic because of Pothole. 4. Thousand Crores of Good Wasted Due to Potholes Pothole the Biggest Problem . They are hurting Everyone knows what the hell potholes are. It’s one the biggest problem faced by every individual in the country. Whatever the way of commute citizen has they have to suffer because. of Potholes. Pothole problem is so common in India and its being there for such a long time that normal people have stop complaining about , people have adopted pothole as the way of life. Damage because of Pothole Society Potholes has caused 3597 last year , which is much more that any other cause of unnatural deaths They create traffic jab wasting lot of human hours Damage Roads leads to burn more fuel leading to more pollution Jerk due to Potholes leads to several health problems including back issues and pain. India Rupees India import 85 Billion Dollar of fuel i.e 6 Lac Core (6,00000,00,00,000). 10% of fuels is wasted due to potholes. We can save lot of this money and increase the rupee value Damaged to imported vehicles caused due to potholes further increase the dollar bill hitting rupee badly Goods Lots of perishable goods like soft fruits are wasted in huge quantity due to potholes. Vehicles life is reduce and their efficiency decrease due to potholes Percentage of damage goods increase compared to good roads. Why potholes are not fixed If potholes are such big problem why they are not fixed. With huge budget allocated to Municipal corporation the road could be fixed. Major question is why even the potholes to such deadly extends occurs.What’s the quality of roads. Is the road being checked and audited. There are many such questions. Whats the answer then. Lets understand why potholes are there in first place Road are faulty.Material is not good. It should be inspected after building. It should check check regularly to find the first occurrence of potholes Even a small pot holes is detected it should be fixed up as early as possible. But due to vast network of road many of the potholes is not noticed by authority and with the course of time it becomes to big. Pothole fixing is not up-to the mark.When a pothole is fixed authority should monitor first defect in fixing , so that authorities can know how good was the road pothole fixing. They can easily identify the good and bad contractor. Saving lot of taxpayers money. Conclusion for Pothole Problem Conclusion is pretty simple monitoring the main issue. If road potholes and new road are properly monitored and fixed on time many bad things like death , back pain , fuel wastage can be avoided. Over a period of time bad contractors and good contractors can be modified , corruption can be fixed and we all will have good roads like any other foreign countries. Monitoring Roads with Humans is not possible, since its not accurate and it may lead to huge amount of corruption which is again a big problem of India. Only way out is to have a good automated system which will monitor the potholes in road and will inform the authority and will keep a log of data for future audit.We called it Pothole Detection System Solution to Pothole Problem in India “Monitoring road with a Fully Automated Drone with Artificial Intelligence (AI) and Deep Learning Algorithms.” i.e Pothole Detection System Above system can solve one of the big problems of our country i.e Pothole. Eventually it will reduce fuel consumption , reduce road accidents , reduce traffic and save man hours. Detail Explanation Road and Pothole Detection System will require 3 Different System Drone Web Application for Interaction Artificial Intelligence Drone for Pothole Detection System A road will be decided for monitoring. The map of road will be feeded into the road. Road may be of 50 to 100 KM or More. Drone will navigate the road in regular interval of time. Navigation Timing of Drone Navigation by Drone can be timely base i.e they will monitor the road every 7 Week. Drone can be triggered on the completion of Event. Event like a Pothole is fixed , Pothole Contract is completed , Road Maintenance Contract is completed etc. Drone will travel as per the feed map and will record videos. Videos will be feeded to the Artificial Intelligence and Deep Learning Layer. This Deep Learning system will identify the potholes. As soon as pothole is detected, Drone will record the geo coordinates of the pothole location and date and time of detection. Details of pothole along with geo coordinates will be sent to Back End , which can be used to authority.Authority can view all the potholes detected by AI System. Drone will also be fitted with laser measurement system. With the laser system the depth of severity of the potholes , its dimension can be known. It will help to define the severity and priority of the Pothole. It will help authority to act on orderly fashion and get rid of more dangerous pothole with highest priority. Machine Intelligence and Deep Learning Layer Machine Intelligence and Deep Learning Layer will help help to find the path hole with Image Processing and Feature extraction. Here we are using two branches of Artificial Intelligence. Deep Learning Deep Learning will help to extract the features from image and video and will help to identify the potholes. This Deep
Machine Learning is an idea to learn from examples and experience, without being explicitly programmed. Instead of writing code, you feed data to the generic algorithm, and it builds logic based on the datasets given. In this article, we are going to cover a simple example of Machine Learning by integrating “Google Video-Intelligence” API with Node.js application. What is Google Video-Intelligence? Google under the project of Machine Learning has introduced Video-Intelligence. This makes video searchable and discoverable by extracting contents of a video with an easy to use REST API. furthermore, you can read more about google video intelligence on the google page. Prerequisites for Machine Learning Basic knowledge of Node.js applications Basic knowledge of Google Cloud Platform Step 1: Enable the Google Video-Intelligence API Sign into your google account with valid user credentials In the Google Cloud Platform Console, go to the Manage resources page and select or create a new project. Make sure that billing is enabled for your project. Also, you can find guidelines on the google docs link. Enable the Cloud Video Intelligence API. Step 2: Authenticating to a Cloud API Service To allow your application code to use a Cloud API, you will need to set up the proper credentials for your application to authenticate its identity to the service and to obtain authorization to perform tasks. The simplest authentication scheme is that of an API key. However, an API key does not allow you to authorize to the service, so it is only usable on public data or data you pass directly to the RPC API. Set up an API key After you enable a Cloud API, Go to “API & Services” through navigation menu. Click “Go to Credentials” to click on the Create Credentials button. Select “API key” from the options. You may wish to copy your key and keep it secure (you can also retrieve it from the API Manager→Credentials page). Set up a service account for machine learning Google Cloud Platform API authentication and authorization (commonly grouped together as “auth”) is typically done using a service account. in addition, A service account allows your code to send application credentials directly to the Cloud API. Go to “API & Services” through navigation menu. Click on “Create Credentials” and select “Service account key”. Select service account as “Compute engine default service account”, select key type as JSON and hit the Create button. Finally, GCP Console will generate a JSON key (as a .json text file), prompt you to download the file to your computer. The generated JSON key will be similar to the following sample JSON key: { “type”: “service_account”, “project_id”: “project-id”, “private_key_id”: “some_number”, “private_key”: “—–BEGIN PRIVATE KEY—–\n…. =\n—–END PRIVATE KEY—–\n”, “client_email”: “<api-name>api@project-id.iam.gserviceaccount.com”, “client_id”: “…”, “auth_uri”: “https://accounts.google.com/o/oauth2/auth”, “token_uri”: “https://accounts.google.com/o/oauth2/token”, “auth_provider_x509_cert_url”: “https://www.googleapis.com/oauth2/v1/certs”, “client_x509_cert_url”: “https://www.googleapis.com/…<api-name>api%40project-id.iam.gserviceaccount.com” } Install and initialize the Cloud SDK Download appropriate google cloud SDK to your OS platform. The installer starts a terminal window and runs the gcloud init command in windows (for other operating systems, manually run the same command). Go through the online guide for detailed information. Provide authentication credentials to your application code by running the following command. Replace [PATH] with the location of the JSON file that contains your credentials. gcloud auth activate-service-account –key-file=[PATH] Obtain an authorization token using the command: gcloud auth print-access-token Copy the access token at somewhere safe. Step 3: Integration with Node.js Application Considering you have already a basic Node.js application ready, we will explain how to use google REST APIs. In order to use video-intelligence, we need to install npm package containing client library. Open the command prompt in the project directory and install google cloud library using the following command npm install –save @google-cloud/video-intelligence Once the package is installed, we will write a code providing local storage video URL to the function. This function will convert the video into the base64 format and pass to the google video intelligence API. Code for Implementation // Imports the Google Cloud Video Intelligence library + Node’s fs library const video = require(‘@google-cloud/video-intelligence’).v1; const fs = require(‘fs’); // Creates a client const client = new video.VideoIntelligenceServiceClient(); /** * TODO(developer): Uncomment the following line before running the sample. */ // const path = ‘Local file to analyze, e.g. ./my-file.mp4’; const path = ‘avengers_trailer.mp4’; // Reads a local video file and converts it to base64 const file = fs.readFileSync(path); const inputContent = file.toString(‘base64’); // Constructs request const request = { inputContent: inputContent, features: [‘LABEL_DETECTION’], }; // Detects labels in a video client .annotateVideo(request) .then(results => { const operation = results[0]; console.log(‘Waiting for operation to complete…’); return operation.promise(); }) .then(results => { // Gets annotations for video const annotations = results[0].annotationResults[0]; const labels = annotations.segmentLabelAnnotations; labels.forEach(label => { console.log(`Label ${label.entity.description} occurs at:`); label.segments.forEach(segment => { let time = segment.segment; if (time.startTimeOffset.seconds === undefined) { time.startTimeOffset.seconds = 0; } if (time.startTimeOffset.nanos === undefined) { time.startTimeOffset.nanos = 0; } if (time.endTimeOffset.seconds === undefined) { time.endTimeOffset.seconds = 0; } if (time.endTimeOffset.nanos === undefined) { time.endTimeOffset.nanos = 0; } console.log( `\tStart: ${time.startTimeOffset.seconds}` + `.${(time.startTimeOffset.nanos / 1e6).toFixed(0)}s` ); console.log( `\tEnd: ${time.endTimeOffset.seconds}.` + `${(time.endTimeOffset.nanos / 1e6).toFixed(0)}s` ); console.log(`\tConfidence: ${segment.confidence}`); }); }); callback(labels); }) .catch(err => { console.error(‘ERROR:’, err); }); After giving the request some time (about a minute, typically), the same request returns annotation results in JSON format. Congratulations! You’ve sent your first request to Cloud Video Intelligence API. Note:- Google Video-Intelligence API is not a free service. You can get free credits for demo purpose. You want any assitance or want to develop you application in Machine Learning you can always contact us here. Resource- Google Video-Intelligence Docs
