Client Requirement

Micro tracer needed redesigning of its existing mobile application for analytical tracer to enable quality assurance in the formula feed industry which can accurately count colored spots on the filter paper-based on selected spot color options.

Manually counting the spot on filter paper cannot provide graphical or statistical results and it was time-consuming. The calculations for percent recovery and mixing quality needed to be accurate. Also, the transfer of input data from the app to the client portal through APIs needed to be implemented.

Spot Counting

Quick Count Testing

This is a single test solution from filter paper. Here filtered paper spots are counted as per app’s selected color, such as RED, BLUE, and RED+ BLUE. The user takes an image in the mobile application of filter paper using a lightbox device.

Validate Mixing Testing

Used multiple testing approach involving three-steps

  • Sample Entry Details
  • Spot Count Sample
  • Test Results for Analysis
MicroTrace Intro & Dashboard
Track Ingredients & Validate Mixing (1)

PROJECT FEATURES AND FUNCTIONALITIES

Detecting Spot

For detecting spot average pixel intensity technique, which creates a mask to distinguish the paper region, it was used. With above created mask, we identified the border pixels and the center of the mask to find the spot’s exact location. Once we got border pixels and the center of the spot, we cropped this spot.

Identifying Color of Spot Detection

We used pixel-to-pixel and histogram analysis. Worked on red and blue color distribution of intensity to check if the colour’s intensity falls in specific range, then we considered blue and red color.

Tracking Ingredients

Added track ingredient feature according to client requirements to save samples for reference.

Sync all data to Micro tracer Portal

Transfer of input data from the App to the client portal was done through APIs.

Spot Detection Using Image Processing In Machine Learning

The spot counting feature was divided into two parts for detecting spot and identifying spot detecting color.

New Entry & Results
Sign-in & Result Summary
LOGIN PAGE

Users can log in to the app; access forgot password/reset password and sign in through Google.

CREATE ACCOUNT

Create an account by adding application details:

LOCATION/ADDRESS

Add location by entering details.

HOME PAGE

VALIDATE MIXING

  • Test multiple filter paper images taken by camera or gallery.

QUICK COUNT

  • Find automatic count spots quickly by clicking filter paper image.

CAPTURE PHOTO

  • The image can be captured by a camera symbol.

RETAKE PHOTO

  • After capturing the image, the camera symbol will change into a retake symbol.

COUNT

  • Offer spot counting of filter paper according to Red, Blue, and Red+Blue colors.
New Entry And Profile
Analytics & Validate Mixing

INSTRUCTION FOR USING VALIDATING MIXING

CREATE NEW ENTRY

Input Tracer information indicates Micro-Tracers product color used, the addition rate into the feed-in question, and sample weight analyzed to calculate expected count for Micro-Tracer particles in the feed sample.

INPUT TRACER COUNT

  • Used for sample series from mixer for analysis
  • Count can be entered manually, or use spot counting
  • Click ‘Calculate’ for ‘Result’ Page
  • Need to analyze ten samples. Tracer Count for minimum of 2 samples is required performing statistical analysis with application

VALIDATE MIXING

There is a listing of valid mixing test.

NEW ENTRY

Create a new test to validate mixing. It comes as Mixer Info, Tracer Info and Input Tracer Count

ADD SAMPLE FROM CAMERA OR GALLERY

Using an icon user can take an image, and the spot will count, and it will add as sample #1. The same process will repeat but minimum 2 samples are required to validate mixing.

RESULTS

Divided into Graph Representation, Calculation and Save Entry

TRACK INGREDIENTS

Serves data entry purpose for Micro-Tracers regarding spot counting

MORE OPTION/HAMBURGER MENU ICON

This option navigates the user to below screens.

  • HOME
  • PORTAL SYNC
  • MY ACCOUNT
  • HELP
  • ABOUT US, CONTACT US & LOGOUT
Add sample from camera

Solution

Technostacks delivered solutions for this approach by developing an easy and user-friendly mobile application using the trending technology aspects of the Machine Learning platform. The following features and solutions were facilitated to achieve accurate analysis results.

Technologies Used

iOS app
android
machine learning
IoT