CN105427321A - Method for detecting existence of soft bag on tray - Google Patents

Abstract

The invention relates to a method for detecting whether a soft bag exists on a tray or not. The method comprises: acquiring an image of the tray from which the soft bag has been discharged; converting the acquired tray image from an RGB color space to an HIS color space; according to the color of an identifier pasted on the soft bag, determining an interested color range; creating a color table three-dimensional array which represents [H] [S] [I], storing the interested color range by using the color table array, and obtaining a final color table; dividing the tray image into a plurality of regions, and for each region, debugging parameters of the region by utilizing the color table to obtain a binary image; eliminating noisy points in the binary image through open operation, extracting communicated regions, and obtaining an area of each communicated region; and comparing the area of each communicated region with a judgment threshold, and if the area of the communicated region, greater than the judgment threshold, exists, indicating that the soft bag is left on the tray. According to the method, whether the soft bag is left on the tray or not can be detected on line.

Description

A kind ofly detect the method whether pallet having soft bag

Technical field

The invention belongs to Digital Image Processing, relate to a kind of method detecting soft bag and whether exist.

Background technology

In the process of producing, due to needs high temperature sterilization, then through the series of steps such as drenching with rain, the pallet filling soft bag is there will be the problem that soft bag leaves over discharging, and owing to being the production of medicine, management and control is very strict.At present in order to get rid of this problem, adopt manual detection.But there is the reasons such as tired and subjective factors interference in manual detection, manual detection exists certain problem.

Summary of the invention

The object of this invention is to provide and a kind ofly can be implemented in the soft bag that line detects on pallet fast and whether there is the method left over.Technical scheme of the present invention is as follows:

Detect the method whether pallet having soft bag, comprise the following steps:

(1) the pallet image of having discharged soft bag is gathered;

(2) transforming the pallet image obtained, is HIS color space by the pallet image collected by RGB color space conversion;

(3) according to the color of the mark that soft bag pastes, determine between interested chromatic zones;

(4) create color table three-dimensional array, three-dimensional array represents respectively [H] [S] [I], with between the interested chromatic zones of this color table storage of array, obtain final color table;

(5) pallet image is divided into multiple region, for each region, the color table utilizing (4) to obtain, debugs the parameter in this region respectively, obtains binary map;

(6) carry out the noise removed by opening operation in binary map, and carry out the extraction of connected region, obtain the area of each connected region;

(7) comparing by each connected region area and decision threshold, if there is the connected region area being greater than decision threshold, then pallet exists soft bag and leave over.

The method step is simple, can better and detect the soft bag leftover problem of pallet fast.

Accompanying drawing explanation

The former figure of Fig. 1

The bianry image that Fig. 2 obtains

Fig. 3 contains little noise

Fig. 4 removes the figure after noise

Embodiment

Below in conjunction with drawings and Examples, the present invention will be described.

(1) color images

Because the Color perception of color space HIS and human eye matches, it is particularly useful in occasions that some illuminations are uneven, and because tone and highlighted, shade have nothing to do, the object of tone to differentiation different colours is very effective.

First, being transformed by the image as Fig. 1 obtained, is HIS color space by the image collected by RGB color space conversion: adopt geometry derivation, according to formula is:

$H=\left\{\begin{array}{cc}\mathrm{\&theta;},& G\&GreaterEqual;B\\ 2\mathrm{\&pi;}-\mathrm{\&theta;},& G<B\end{array}\right.,where\mathrm{\&theta;}={\cos }^{-1}\left(\frac{(R-G)+(R-B)}{2*\sqrt{{(R-G)}^2+(R-B)*(G-B)}}\right)$

$S=1-\frac{3*min(R,G,B)}{R+G+B}$

$I=\frac{R+G+B}{3}$

Secondly, create color table array ColorTable [361] [101] [256], three-dimensional array represents respectively [H] [S] [I], with between our interested chromatic zones of this color table storage of array (blueness above soft bag), obtains final color table;

Finally, searching loop whole image, if pixel (X, Y) drops in color table, otherwise then this place's pixel value is set to 1 and is set to 0; Obtain a bianry image, as Fig. 2;

Remarks: because pallet dimension is 1700mm*1700mm, adopt on-the-spot natural light during detection, so in order to ensure the stability detected, we have divided six regions on image, to the independent regulating parameter in each zonule.

(2) connected region is extracted and counting

Adopt 4 to face territory region growing approach to bianry image and extract each connected region.Be implemented as follows:

Searching loop bianry image, when image (x, y) place pixel is 1, region growing is carried out as seed by this point, and statistical pixel number, after this feed search completes, its value is set to 0, continue search by new seed, until not regrowth, determine connected region 1.

Searching loop bianry image, searching for another place's pixel value is that the point of 1 is as seed determination connected region 2.

Circulate successively ... .., stopping stop condition is: then can grow without point;

Acquire each connected region, demonstrate the area of each connected region final.

(3) with or without judgement

According to Fig. 4, if there is the area of any one connected region to be greater than the decision threshold of setting, then there is soft bag above pallet and leave over.

Claims (1)

1. detect the method whether pallet having soft bag, comprise the following steps:

(1) the pallet image of having discharged soft bag is gathered;

(2) transforming the pallet image obtained, is HIS color space by the pallet image collected by RGB color space conversion;

(3) according to the color of the mark that soft bag pastes, determine between interested chromatic zones;

(4) create color table three-dimensional array, three-dimensional array represents respectively [H] [S] [I], with between the interested chromatic zones of this color table storage of array, obtain final color table;

(5) pallet image is divided into multiple region, for each region, the color table utilizing (4) to obtain, debugs the parameter in this region respectively, obtains binary map;

(6) carry out the noise removed by opening operation in binary map, and carry out the extraction of connected region, obtain the area of each connected region;

(7) comparing by each connected region area and decision threshold, if there is the connected region area being greater than decision threshold, then pallet exists soft bag and leave over.