This section describes AeroCapture functionalities
For every different type of application, you can create a project. For example, if you are having application of transmission tower inspection, then you can create a project for that, if you are having application of solar power infra inspection you can create a new separate project for that.
You have to enter the name of the project and the client’s name, so it helps in better understanding in future whenever you need to access the same.
Create a New Project
Project Name: Project Name can be in alphabets and using numbers as well.
Client Name: Here you enter the name of the client for whom the project has been run.
All Projects: Once projects are created, you see list of projects. Shows Total count of the Project that have been created.
All Created Projects are Shown here
Enables to Delete a Project
Delete Project: In case of deleting the project, you can do that by clicking on the three dots and directly remove the project. This action is irrecoverable.
Here you need to enter plan name and description of the same. You can create any number of plan under a project. There can be one plan for each asset or group of assets to inspect. For example, if you want to inspect all transmission towers of a city, then you can create one plan for each transmission tower or one plan for all towers of nearby geo-graphical area.
Plan Name: Plan name can be the mission behind the plan.
Description: Description consists of the purpose of the project. (Example; Transmission Tower Inspection, Road Monitoring, Solar Panel Inspection, Parking Space Management Etc.)
Create a New Plan
One plan is created, then you see all created plans.
Delete: In case of deleting the project, you can do that by clicking on the three dots and directly remove the project. This action is irrecoverable.
Enables you to Delete a Created Plan
Image-Set and Video Set
You need to create an image set to upload all images into a specific folder / set. As shown below, you need to define name to the image set.
This enables you to upload images as per your tasks to be performed.
Create ImageSet for Analytics
Upload Image Sets:
You can add and delete number of image/video sets on which the task is to be performed. There are four supported methods to get the images for analytics.
Upload from Computer: This directs you to pick image files from your local device. Once you upload the files would be reflected on your screen.
Import from Video: AeroCapture supports importing of images from a video. You can input a video as source and based on number of frames per second setting, it can convert the video into images. You can input video from your local device.
Import from Projects: AeroCapture supports importing images from projects created in AeroMegh account. When flight planning is done in AeroPlanner or AeroGCS, that time project is created. The automated image transferring from drone to AeroCapture when drone is having AeroStack installed is supported by AeroCapture. Therefor all images stored in a project while drone is flying with AeroStack can be imported into AeroCapture for automated analytics. This makes entire process automated and helps to get real time AI results while drone is flying.
Import from AeroStream: AeroCapture supports importing of images from a video captured using AeroStream service. This enables you to pick images from your AeroStream account. You can input video as source and based on number of frames per second, it can convert the video into images. If you have AeroStream account and videos are recorded in AeroStream, then those videos can be used to generate images based on frame setting. Following image shows all options of importing images into AeroCapture.
Upload or Import ImageSets for Analytics
Upload from Computer: This directs you to pick files from your local device. Once you upload all the files would be reflected on your screen.
Allows to Import ImageSets from your Device
Import from Project: This enables you to pick images from your AeroStream account.
Import from Video: This directs you to pick video from your local device. You can browse a video on your local computer and select images per second to extract from a video.
Allows you to extract images from a uploaded video
Import from AeroStream: This enables you to pick images from your AeroStream account. This directs you to pick video from AeroStream account of AeroMegh. You can browse all video of your AeroStream account and select any video you want to extract to convert to images. Post that select images per second to extract from a AeroStream video.
Displaying all ImageSets and VideoSets
Analytics tasks is the core component of the AeroCapture service. This task defines the actual analytics you want to do on images or videos. You can select the objects to be detected from image set or video and a report of object detection will be generated by system according to that. There are pre-defined, machine learning trained objects available in AeroCapture which you can select to detect from images or videos. For example, if you are doing transmission tower inspection project, then you can select image set of transmission tower and then select various objects / defects as objects to be detected from transmission tower image set.
Here you have to create a task based on the project that you have to run. It has the following Mandatory fields.
Create A New Task
Image/Video: You have to select if the following task is going to be performed on images sets or video sets.
Enter Title for This Task: Title can be in alphabets and numbers representing the task details.
Select Images from Plan: This enables you to select multiple images/videos from the plan.
Select Image/Video Set: Media can be selected from the device itself or where the data is stored.
Analysis to be Performed: This Allows you to select type of defects, it has number of options like; Helmet, Person, Hotspots, Rusted tower, Pothole, etc. You can select the area of identification through this.
Task Summary: Once all credentials are filled, AeroCapture directs you to task summary. Displaying all the necessary information stored.
Finish: Once you click on “finish”, AeroCapture notifies that the task has been began. As soon as task is stared, the AeroCapture system sends all images and videos to GPU (Graphical Processing Unit) farm along with objects to be detected from all images. Each image is scanned/analyzed by GPU farm against all the objects to be detected from those images. Every success result is stored as analysis of objects from those images. The task of analysis will complete based on volume of images and number of objects to be detected.
Notification pop-up before starting a task
When the task is getting executed by GPU farm, you see AeroCapture shows its ‘pending’. As soon as it is completed, its status will become ‘completed’.
Shows all Tasks Analysis and their status
Reporting of image and video analytics using AI and ML for your application is important aspect of AeroCapture. AeroCapture supports two types of reports.
This report gives you the summary of the task performed. It generates overall objects detected from all given images of video as summary output of task. For example, if you are doing project of transmission tow inspection, and you want to find out total 20 objects as transmission tower component or defects then AeroCapture will generate a graph of summary of all objects found from AI analytics.
Summary report of performed task
Reports are generated once the task has been performed. Report section consists of Name of the Task and Created Date of the same.
See Detailed and Summary Reports
Details report shows all images and all the objects detected from those images. Here you can see that car and parking space are two objects detected from 10 images. Each image shows total objects detected from it and number.
Detailed Report as per uploaded Imageset
Image Particular Report: AeroCapture allows you to check one-to-one image analytics of the images uploaded. Helps you understand better. You can analyze based on each image. Following image shows the analytics where 9 cars are detected from image and total 8 parking space are detected.