Health Care Industry Machine Vision for Pill Quantity Detection

Inspection principle

In traditional counting and packaging machines, tablets or capsules are usually placed in a hopper at the end of the machine, either manually or automatically. The product will be fed from the hopper into the vibrating feeding system and distributed to three sets of stainless steel cascading feed trays, each of which is in turn connected to six slides. The chute vibrating device of the feed tray differentiates the products from each other and moves them forward along the different chutes. As the tablets or capsules leave the end of the last feeding tray, they pass through high-speed infrared optical sensors, and after counting the amount of product in free fall, guide the product into individual containers with the aid of chutes and filling nozzles.

In order to find out whether a medicine bottle is filled with a given number of tablets, the data acquired by the sensor needs to be transmitted to the embedded PC, so that the number of tablets loaded into each medicine bottle can be calculated. Once a vial in the machine is full, the PC will close the cap and eventually suspend the vibratory feeding system associated with it until the other vials are full. The bottles are then driven by the system to proceed, depending on whether they pass the system acceptance or are rejected, depending on whether they are loaded with a given number of tablets.

Inspection details

In order for the machine to detect if tablets are broken or not yet formed, Efit Vision installed a camera above each vibratory feeder in the counting and bottling machines. To illuminate the feed pan, a darkfield lighting system was developed; the system consists of four LED strip lights that project light into the feed pan at a -angle. Darkfield illumination technology ensures that light incident on the mirror surface of the stainless steel feed pan is reflected away from the camera, so that an image of the matte surface of the tablet or capsule can be efficiently obtained.

The image data of the tablets or capsules in the feeding tray are transmitted to a dedicated PC through the interface. In the PC, the image is processed by the EMVP image processing software developed by the company. The main purpose of the test is to identify small particles or half-tablets that may be present. In order to achieve this purpose, EMVP software needs to determine the length, width, area size and integrity of the tablet by analyzing the image of the tablet.

A PC-based visual inspection system is connected to an embedded PC that performs counting functions, and communicates in real time. If a tablet is determined by the vision system to be defective, it is still bottled by the machine. However, the PC running the vision system software would then flag the tablet as defective, and would signal a rejection to the embedded PC that was performing the counting function. Next, the embedded PC will detect, based on the reject signal received from the vision system, which bottle contains the reject tablet. Afterwards, vials containing substandard tablets will be removed from the production line after leaving the bottling machine.