1 min. 40 sec.read
Maintenance company, handling 70 ports across the USA, integrates simple, cost-effective solutions to record, store, categorize, and analyze video footage through body-worn cameras on staff. VEER recognized the problem areas and offered solutions resulting in over 10 benefits for the business.
In South America, an important maintenance company was running an electrical network surveillance project. This project in particular included body-worn cameras only. While working offline, at the end of every day 12 TB of data is stored and copied from each terminal out of 70 terminals handled.
The situation was as follows:
– There were 1000 team members distributed all over the country to grant full coverage.
– Each team member wore one camera so in total there were 1000 body-worn cameras distributed to team members.
– Every day, from 60 to 70 terminals were ready for daily data upload.
– More than 50 employees were allocated to watch the footage live.
– There were up to 20 automatic video categorization tags.
– Video categorization and classification were done manually.
– 2 to 3% only of the full videos were analyzed.
– Recording was done offline and footage analysis was done offline.
The maintenance company had huge amounts of data influx every day that did not receive proper analysis and/or classification. They were losing great amounts of time, personnel, and resources.
VEER recognized the needs and shortcomings such as relying on humans to analyze great amounts of data. One person can only watch one monitor while a computer is able to analyze several streams of footage simultaneously. Another challenge faced was that this software was on-demand. VEER built the Body-worn cameras software fully. As it was not an off the shelf solution, it was developed to perfection through machine learning to match expectations.
– The company decided to automate the project by using the Object Classifier Module by VEER through body-worn cameras.
– The solutions and advantages gained were as such:
– Reduced the number of employees, allocated to watch the footage live, from more than 50 to 2 employees only.
– Video analysis coverage jumped from only 3% to 100% coverage and reduced human error to almost 0%.
– Video categorization and classification became automatic.
– Each video uploaded was broken into several location tags to track employee movements throughout each day.
(i.e. 70% on roof – 25% highway – 5% office)
– Reduction of the number of terminals from 70 to only 2 terminals for processing and a storage server.
– Reduction of storage automatically by storing needed videos only.
– Possibility of future installation of live analytics for the body-worn cameras on demand.