FMCG Packaging Using AI-Based Machine Vision System

Shriyansh Tiwari, Qualitas Technologies

The fast-moving consumer goods (FMCG) industry is not only incredibly complex – with a number of systems and value chains involved – but it is also ripe with many opportunities for innovation. The new industrial age has witnessed the rise of the Internet of Things (IoT) which has enabled physical systems to communicate data and automatically convert the insights extracted from this data into actions. This is actually where machine vision steps in.

In simple words, machine vision is a combination of hardware and software that helps machines “see” something, extract relevant pieces of information and then act accordingly. Given the amazing range of benefits that machine vision offers, its demand has surged explosively across all major industries. The global machine vision market which was valued at 12.29 billion USD in 2020, is poised to expand at a CAGR of 6.9%from 2021 to 2028.

Given the huge gamut of processes and operations within the FMCG industries, ranging from production and processing to packaging and distribution, machine vision can significantly boost the safety and quality of products across the numerous stages of the value chain. Let’s see how machine vision has impacted the FMCG packaging segment using a case study.

100 Percent Accuracy with Machine Vision in FMCG Packaging – Case Study

Our client, one of the largest snack companies in the world, offers products like biscuits, gum, chocolate, beverages, etc. Initially, our client had additional manual labor in place for manual inspection. High production speeds led to labor fatigue and didn’t lead to any better packaging quality control. Some critical challenges being faced by our client were:

Our client needed a vision solution to inspect the following parameters on primary and secondary packaging:

  1. The integrity of the OCR printed on primary packing (speed – 140 parts per minute).
  2. Identify flap defects in primary packing (speed – 140 parts per minute)
  3. Flap gaps (>5mm)
  4. Flap misalignment (>1mm)
  5. Glue gap (>2.5mm)
  6. Presence/absence of secondary packaging in the infeed bucket.
  7. Segregation of SKUs based on their variants

We at Qualitas Technologies followed a deliberate step-by-step process to come up with a practical vision solution meeting our client’s demands. Firstly, the camera and apt lighting systems were put into place to best capture the defects. The solution was then trained on the acquired images of primary and secondary boxes, and all kinds of defects. AI-based object detection, anomaly detection, OCR, and classification techniques were utilized to carry out all kinds of inspections. During the course of its performance, DL programs will help improve the vision solution.

With our vision solution in place, all the defects were identified correctly, and the action of rejection was performed on the defective ones. We observed that:

Implications of SKU Mismatch and Incomplete Products

Packaging lines are one of the most complex areas in FMCG industries. Sometimes, even getting the right product in the right package can feel like a challenge. With a number of packaging combinations being used and different kinds of products being churned out simultaneously, mistakes become naturally probable, especially due to human error. However, even small errors such as mislabeling or mismatched products can have serious repercussions. Some critical implications of not ensuring quality packaging are the following:

Trivial errors such as mismatches and mislabeling can cause friction in the supply chain. With such intense competition in the FMCG sector, customer dissatisfaction can in no time translate into the loss of customers. A more serious consequence of such a mismatch is that some packagescould enter the supply chain with incorrect allergen information. If packaging errors put the consumers’ lives at risk, they must be avoided at all costs.

The product recall frequency reflects the consequences of inaccurate inspection. Ultimately, the products that don’t meet the set quality standards are recalled, leading to the associated brand, monetary, and waste problems.

Dissatisfied customers have the right to approach consumer forums to voice their concerns. The consumer forum can then reprimand the product’s company with strict actions or fines. In one such case, the Mumbai Suburban District Consumer Forum directed Parle Products to pay a compensation of Rs 60,000to the complainant who had found a piece of sellotape inside a Parle Monaco biscuit packet costing Rs 7.

How Machine Vision Benefits FMCG Packaging Industries

Considering the nature of tasks in FMCG packaging industries that are highly error-prone but simultaneously demand strict compliance, solely relying on manual labor is naturally not a wise choice. Automation of such complex tasks with machine vision is an effective way to beat the aforementioned challenges. Some clear benefits of machine vision automation are the following:

In contrast to the vast scope of errors associated with manual labor, machine vision systems can perform SKU mismatch identification and object detection with almost 100% accuracy, ensuring optimal quality control of packaging material.

Automation of packaging with machine vision ensures a seamless workflow and promises a quick turnaround.

Higher accuracy ensures quality packaging, which results in lower recall rates and enhanced customer satisfaction.

By cutting down the number of laborersrequired, companies can considerably reduce their operational and maintenance expenses in the long run.

Conclusion

This article delved into the range of applications for machine vision systems in the FMCG sector, particularly the packaging stage. Using a case study, we understood the challenges a key player faces in the FMCG segment and how Qualitas Technologies came up with an effective solution to address the same. Moving further, we understood the implications of some packaging errors and how machine vision technology helps industries in the FMCG segment practice maximal quality control of packaging material.