DD292976A5 - METHOD FOR THE ANALYSIS OF SLAUGHTER BODY AGENTS BY IMAGE PROCESSING - Google Patents

Abstract

The invention relates to a method for the analysis of slaughtered animal body parts by means of image processing as a prerequisite for the fully automatic coarse dissection, classification and acquisition of exact morphological data in the meat processing industry or in slaughterhouses. The invention creates a condition for rationalizing physically heavy and inexact work. According to the invention, both the outer contour and the layer of bacon as well as the entire vertebral column are identified. The intervertebral layers are the prerequisite for the definition of the section for the dissection. The outer contour, the layer of bacon and the exact position of the sixth lumbar vertebra are used to calculate the classification. Due to the focus of the determined vortices, the method simultaneously provides parameters for exact morphological examinations {fully automatic coarse decomposition; classification; Schlachttierkoerperhaelften; Analysis of bone structure; Meat cutting; Contour; spine; Image processing}

Description

For this 2 pages drawings

Field of application of the invention

The invention relates to a method for analyzing carcass halves by means of image processing in the meat processing industry or in slaughterhouses.

Characteristic of the known technical solutions From US-PS 3,916,484 a method is known in which the cutting positions of the carcass half by color Marked by professionals and scanned with a video camera. By an evaluation electronics

the signals are used to automatically control the decomposition.

Due to the need of a specialist who must apply the color marking on the carcasses, is not fully automatic

and given an unrelated solution.

A method and an apparatus has become known by US-PS 4,662,029, wherein for each section at the

Carcass half another camera is used, the image compared with a computer present in the pattern

becomes. A reference point is located for each cut and, after pattern matching, the control parameters for the decomposition are transferred. This solution requires a great deal of material and only takes into account the outer contour of the carcasses.

Only a limited accuracy in the decomposition can be guaranteed. A method for classifying a carcass, in particular a half pig, in commercial classes is by the DE-PS 2,946,912 become known. Here are video-optical devices outer contour features as a silhouette

recorded, superimposed on a measuring grid and determined therefrom the commercial class of the carcasses. The maximum waist width and minimum waist width and an angle α are detected by applying a tangent to the middle region of the thigh.

Since only the outer contour is determined, only a limited accuracy is feasible. A method for non-contact determination of quality characteristics of a test object of the meat category is

become known from DE-OS 3,047,490. In this case, the test object is detected by a video camera and the meat or fat tissue is distinguished depending on the relative brightness.

The image of the sample is digitized with a discrimination-typical threshold. This procedure is very elaborate and

does not provide sufficient accuracies for determining quality characteristics.

A method and apparatus for classifying meat has become known from DE 2,728,913. This takes place

an electro-optical measurement of several parameters for classification. The parameters are detected with a video camera and evaluated in a microprocessor system.

It is the ham contour, the ham side surfaces to the separating vortex, the maximum ham thickness, the fat thickness,

the meat-bacon-ratio and the weight determined.

However, this method can not use the entire bone structure for dissection. Object of the invention

The aim of the invention is to determine the required cutting and quality parameters contactless and fully automatic with a high accuracy for a decomposition and classification by an exact analysis of slaughter animal body halves.

Explanation of the essence of the invention

The invention has for its object to determine by inclusion of the bone structure for coarse dissection accurate positional parameters of the cutting paths, to determine the lean meat content in the lumbar region and the bacon thickness for classification and to allow accurate morphological studies.

The solution according to the invention consists of known image processing systems and comprises the following method steps:

1. A camera photograph of the carcass half is detected so that the scene includes the spine, including the sacrum, as well as the outer contours of the carcass. The spine is approximately parallel to the image horizon.

2. An analysis of the spine taking into account the outer contours of the animal body and the different gray levels of meat, fat and bone is achieved by a vertical evaluation of the image matrix.

3. Through the location of the spine found to the outer contours of the carcass and a vertical gray value analysis, which cuts the contours and the spine, a decision can be made throughout the scene on the presence of a left or right halon animal body.

4. A picture is taken containing the spine in the lumbar region. Due to the position of the sacrum to the spine at an obtuse angle this leaves the picture.

5. Taking into account the angle formed by the spine and sacrum, may be due to a functional

Grayscale determination within the image section a fixed point are determined whose position is determined by the steepest gradient of the functional gray value determination.

6. The image section is filtered with horizontal gradient operators and binarized so that the inter-vertebral layers receive maximum contrast with their surroundings.

7. In the image section, all intervertebral layers are identified and their position determined.

8. Taking into account the parameters already determined and the intervertebral layers found, the entire spinal column is examined and the position of all intervertebral layers determined.

9. From the determined positions, the cutting parameters for the coarse decomposition are derived.

10. According to the 6th lumbar vertebrae, the lean meat content and the bacon thickness for the classification are determined by vertical evaluation of the gray values.

11. Exact morphological examinations can be carried out from all parameters, depending on the imaging scale. With the solution according to the invention, a complete analysis of the spinal column from sacral bone to atlas is carried out with image processing, and from this, cutting parameters for the decomposition and parameters for the classification are determined.

exemplary

The essence of the invention is explained in more detail by the following example:

Fig. 1: shows the part of a pig half in the lumbar region. This picture is unprocessed and lies with 64 gray values in the

Calculator before, Fig. 2: shows a processed representation of Figure 1. This image shows the analyzed vertebrae and intervertebral layers whose positional parameters are present in the computer.

As a first step after image acquisition according to Figure 1, the outer contour and the contour between the bacon layer and meat layer are detected.

¹ P Dependency of these parameters defines a picture section that contains the spinal column.

By a vertical gray value analysis, the exact position of the spine is determined.

The vertical gray value analysis, which intersects the contours and contains the information about the position of the spine, provides the decision as to whether it is a left or a right half of the carcass.

An image detail in FIG. 2, which contains the lumbar vertebrae, shows that the sacrum forms an obtuse angle to the spine and leaves the image detail.

The fixed point, which is characterized by the 45 ° straight line in FIG. 2, is determined by a functional gray value determination within the image detail. The fixed point position is determined by the steepest gradient of this gray scale determination.

The image section is filtered with a horizontal gradient operator, the strength of which is determined by the gray value average of the entire image section. The greyscale also determines the subsequent binarization threshold.

Through these operators, the intervertebral layers represent clear signals, which are obtained by vertical gray scale determination.

An iterative search for local maximas of the vertical grayscale determination provides the positions of the existing ones

Intervertebral layers. In this case, a suppression is realized on the left and right of a found intervertebral layer, in order to cover existing interference. The positions of the intervertebral layers, together with the positional parameter of the spinal column, form the starting points for a contour tracing, so that the whirls filled in white in FIG. 2 are formed.

Based on the determined parameters, the entire carcass half and all intervertebral layers are analyzed.

From the position of the spine and the Zwischenwir coat the cutting parameters for the coarse decomposition are calculated

and thus control the decomposition.

The 6th lumbar vertebra serves as a starting point for the determination of the lean meat content by means of a circumscribing rectangle.

Furthermore, the bacon thickness is detected. Both quantities are charged to a classification parameter in a known manner.

The focal points of the vertebral contours form the basis for a mathematical modeling as well as for an exact morphological examination.

Claims (3)

1. A method for analyzing carcass halves by means of image processing, in particular of Pork halves, which are determined by Bildverarbeitungshard- and software outer contour, bacon layer and a FleicJh-back fat ratio, characterized in that first of the carcass half of a spine with all intervertebral layers with known video technology is performed and as a fixture for dissection and classification of the sacral spine is assumed. 2. The method according to claim 1, characterized in that are determined from the position of the intervertebral layers cutting parameters for the decomposition. 3. The method according to claim 1, characterized in that are determined from the location of the sixth lumbar vertebra and the outer contour and the bacon layer bacon and meat dimensions for the classification.